item_id
stringlengths
21
26
family
stringclasses
383 values
twin_of
stringlengths
0
21
system
stringlengths
1.51k
1.51k
user
stringlengths
18.6k
288k
input_tokens
int64
6.38k
123k
budget_tokens
int64
1k
3.07k
cut
int64
19
151
steps
int64
32
166
resolved
bool
1 class
image
stringlengths
25
33
source
stringclasses
1 value
facts
stringlengths
585
1.33k
probes
stringlengths
1.29k
1.44k
ctx_compress-s1-00655
shipmentworks
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] restock_warehouse drops override values of 0, False and empty string. It lives in src/shipmentworks/limits.py; `python -m pytest tests/test_limits.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the shipmentworks la...
97,523
3,072
114
123
false
synth:7985012260870010328@1.8481
synth
[{"aliases": ["src/shipmentworks/limits.py", "limits.py"], "kind": "file", "value": "src/shipmentworks/limits.py"}, {"aliases": ["tests/test_limits.py", "test_limits.py"], "kind": "file", "value": "tests/test_limits.py"}, {"aliases": ["src/shipmentworks/core.py", "core.py"], "kind": "file", "value": "src/shipmentworks/...
[{"answers": ["src/shipmentworks/limits.py", "limits.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent'...
ctx_compress-s1-00655-twin
shipmentworks
ctx_compress-s1-00655
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] restock_warehouse drops override values of 0, False and empty string. It lives in src/shipmentworks/limits.py; `python -m pytest tests/test_limits.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the shipmentworks la...
42,792
2,139
59
123
false
synth:7985012260870010328@1.8481
synth
[{"aliases": ["src/shipmentworks/limits.py", "limits.py"], "kind": "file", "value": "src/shipmentworks/limits.py"}, {"aliases": ["tests/test_limits.py", "test_limits.py"], "kind": "file", "value": "tests/test_limits.py"}, {"aliases": ["src/shipmentworks/core.py", "core.py"], "kind": "file", "value": "src/shipmentworks/...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["changelog.md", "docs/changelog.md"], ...
ctx_compress-s1-00656
palletdesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] age_bin leaves a trailing dash when the title ends in punctuation. It lives in src/palletdesk/rules.py; `python -m pytest tests/test_rules.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the palletdesk layout. [tool...
93,885
3,072
92
111
false
synth:8963458370904675212@1.63177
synth
[{"aliases": ["src/palletdesk/rules.py", "rules.py"], "kind": "file", "value": "src/palletdesk/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["data/sku_5.csv", "sku_5.csv"], "kind": "file", "value": "data/sku_5.csv"}, {"aliases": ["data/sh...
[{"answers": ["src/palletdesk/rules.py", "rules.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's mos...
ctx_compress-s1-00656-twin
palletdesk
ctx_compress-s1-00656
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] age_bin leaves a trailing dash when the title ends in punctuation. It lives in src/palletdesk/rules.py; `python -m pytest tests/test_rules.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the palletdesk layout. [tool...
49,658
2,482
56
111
false
synth:8963458370904675212@1.63177
synth
[{"aliases": ["src/palletdesk/rules.py", "rules.py"], "kind": "file", "value": "src/palletdesk/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["data/sku_5.csv", "sku_5.csv"], "kind": "file", "value": "data/sku_5.csv"}, {"aliases": ["data/sh...
[{"answers": ["src/palletdesk/rules.py", "rules.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's mos...
ctx_compress-s1-00657
policyforge
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] mint_grant double counts carried amounts when the period closes on a weekend. It lives in src/policyforge/cli.py; `python -m pytest tests/test_cli.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the policyforge layo...
69,926
3,072
114
137
false
synth:37675748760295411@2.14645
synth
[{"aliases": ["src/policyforge/cli.py", "cli.py"], "kind": "file", "value": "src/policyforge/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["data/principal_0.csv", "principal_0.csv"], "kind": "file", "value": "data/principal_0.csv"}, {"aliases": [...
[{"answers": ["src/policyforge/cli.py", "cli.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most r...
ctx_compress-s1-00657-twin
policyforge
ctx_compress-s1-00657
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] mint_grant double counts carried amounts when the period closes on a weekend. It lives in src/policyforge/cli.py; `python -m pytest tests/test_cli.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the policyforge layo...
47,472
2,373
75
137
false
synth:37675748760295411@2.14645
synth
[{"aliases": ["src/policyforge/cli.py", "cli.py"], "kind": "file", "value": "src/policyforge/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["data/principal_0.csv", "principal_0.csv"], "kind": "file", "value": "data/principal_0.csv"}, {"aliases": [...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["cli.py", "src/policyforge/cli.py"], "...
ctx_compress-s1-00658
binkit
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] reserve_sku raises WarehouseRangeError for bytes input in strict mode instead of decoding it. It lives in src/binkit/rules.py; `python -m pytest tests/test_rules.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the b...
46,366
2,318
91
104
false
synth:4700481703176994098@1
synth
[{"aliases": ["src/binkit/rules.py", "rules.py"], "kind": "file", "value": "src/binkit/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/binkit/store.py", "store.py"], "kind": "file", "value": "src/binkit/store.py"}, {"aliases": ["src/bi...
[{"answers": ["src/binkit/rules.py", "rules.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most re...
ctx_compress-s1-00658-twin
binkit
ctx_compress-s1-00658
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] reserve_sku raises WarehouseRangeError for bytes input in strict mode instead of decoding it. It lives in src/binkit/rules.py; `python -m pytest tests/test_rules.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the b...
22,693
1,134
47
104
false
synth:4700481703176994098@1
synth
[{"aliases": ["src/binkit/rules.py", "rules.py"], "kind": "file", "value": "src/binkit/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/binkit/store.py", "store.py"], "kind": "file", "value": "src/binkit/store.py"}, {"aliases": ["src/bi...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["rules.py", "src/binkit/rules.py"], "k...
ctx_compress-s1-00659
calendardesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] rotate_shift returns the upper bound for every value instead of clamping into the range. It lives in src/calendardesk/limits.py; `python -m pytest tests/test_limits.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping th...
72,915
3,072
104
121
false
synth:3898880966729295318@1.5587
synth
[{"aliases": ["src/calendardesk/limits.py", "limits.py"], "kind": "file", "value": "src/calendardesk/limits.py"}, {"aliases": ["tests/test_limits.py", "test_limits.py"], "kind": "file", "value": "tests/test_limits.py"}, {"aliases": ["data/shift_6.csv", "shift_6.csv"], "kind": "file", "value": "data/shift_6.csv"}, {"ali...
[{"answers": ["src/calendardesk/limits.py", "limits.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's...
ctx_compress-s1-00659-twin
calendardesk
ctx_compress-s1-00659
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] rotate_shift returns the upper bound for every value instead of clamping into the range. It lives in src/calendardesk/limits.py; `python -m pytest tests/test_limits.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping th...
51,594
2,579
72
121
false
synth:3898880966729295318@1.5587
synth
[{"aliases": ["src/calendardesk/limits.py", "limits.py"], "kind": "file", "value": "src/calendardesk/limits.py"}, {"aliases": ["tests/test_limits.py", "test_limits.py"], "kind": "file", "value": "tests/test_limits.py"}, {"aliases": ["data/shift_6.csv", "shift_6.csv"], "kind": "file", "value": "data/shift_6.csv"}, {"ali...
[{"answers": ["src/calendardesk/limits.py", "limits.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's...
ctx_compress-s1-00660
captionflow
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] hyphenate_chunk raises LexemeBoundError for bytes input in strict mode instead of decoding it. It lives in src/captionflow/compat.py; `python -m pytest tests/test_compat.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mappi...
77,355
3,072
96
111
false
synth:9078019413464333084@1.61295
synth
[{"aliases": ["src/captionflow/compat.py", "compat.py"], "kind": "file", "value": "src/captionflow/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["src/captionflow/rules.py", "rules.py"], "kind": "file", "value": "src/captionflow/rules....
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["chunk_3.csv", "data/chunk_3.csv"], "k...
ctx_compress-s1-00660-twin
captionflow
ctx_compress-s1-00660
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] hyphenate_chunk raises LexemeBoundError for bytes input in strict mode instead of decoding it. It lives in src/captionflow/compat.py; `python -m pytest tests/test_compat.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mappi...
51,123
2,556
66
111
false
synth:9078019413464333084@1.61295
synth
[{"aliases": ["src/captionflow/compat.py", "compat.py"], "kind": "file", "value": "src/captionflow/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["src/captionflow/rules.py", "rules.py"], "kind": "file", "value": "src/captionflow/rules....
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["src/captionflow/tables.py", "tables.p...
ctx_compress-s1-00661
tileledger
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] project_fix_window returns one item too few on every page. It lives in src/tileledger/audit.py; `python -m pytest tests/test_audit.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the tileledger layout. [tool_use kin...
32,304
1,615
69
82
false
synth:6222453055454553365@1
synth
[{"aliases": ["src/tileledger/audit.py", "audit.py"], "kind": "file", "value": "src/tileledger/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["src/tileledger/cli.py", "cli.py"], "kind": "file", "value": "src/tileledger/cli.py"}, {"aliases"...
[{"answers": ["src/tileledger/audit.py", "audit.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's mos...
ctx_compress-s1-00661-twin
tileledger
ctx_compress-s1-00661
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] project_fix_window returns one item too few on every page. It lives in src/tileledger/audit.py; `python -m pytest tests/test_audit.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the tileledger layout. [tool_use kin...
17,200
1,000
43
82
false
synth:6222453055454553365@1
synth
[{"aliases": ["src/tileledger/audit.py", "audit.py"], "kind": "file", "value": "src/tileledger/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["src/tileledger/cli.py", "cli.py"], "kind": "file", "value": "src/tileledger/cli.py"}, {"aliases"...
[{"answers": ["src/tileledger/audit.py", "audit.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's mos...
ctx_compress-s1-00662
reorderkit
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] reserve_sku produces doubled dashes for consecutive separators. It lives in src/reorderkit/cli.py; `python -m pytest tests/test_cli.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the reorderkit layout. [tool_use ki...
79,485
3,072
117
144
false
synth:3446627053468698007@2.20887
synth
[{"aliases": ["src/reorderkit/cli.py", "cli.py"], "kind": "file", "value": "src/reorderkit/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["src/reorderkit/report.py", "report.py"], "kind": "file", "value": "src/reorderkit/report.py"}, {"aliases": [...
[{"answers": ["src/reorderkit/cli.py", "cli.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most re...
ctx_compress-s1-00662-twin
reorderkit
ctx_compress-s1-00662
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] reserve_sku produces doubled dashes for consecutive separators. It lives in src/reorderkit/cli.py; `python -m pytest tests/test_cli.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the reorderkit layout. [tool_use ki...
50,306
2,515
77
144
false
synth:3446627053468698007@2.20887
synth
[{"aliases": ["src/reorderkit/cli.py", "cli.py"], "kind": "file", "value": "src/reorderkit/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["src/reorderkit/report.py", "report.py"], "kind": "file", "value": "src/reorderkit/report.py"}, {"aliases": [...
[{"answers": ["src/reorderkit/cli.py", "cli.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's most re...
ctx_compress-s1-00663
policykit
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] extend_grant drops override values of 0, False and empty string. It lives in src/policykit/format.py; `python -m pytest tests/test_format.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the policykit layout. [tool_u...
57,315
2,865
78
97
false
synth:3000078996865310840@1.19705
synth
[{"aliases": ["src/policykit/format.py", "format.py"], "kind": "file", "value": "src/policykit/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["src/policykit/limits.py", "limits.py"], "kind": "file", "value": "src/policykit/limits.py"},...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["claim_0.csv", "data/claim_0.csv"], "k...
ctx_compress-s1-00663-twin
policykit
ctx_compress-s1-00663
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] extend_grant drops override values of 0, False and empty string. It lives in src/policykit/format.py; `python -m pytest tests/test_format.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the policykit layout. [tool_u...
35,403
1,770
54
97
false
synth:3000078996865310840@1.19705
synth
[{"aliases": ["src/policykit/format.py", "format.py"], "kind": "file", "value": "src/policykit/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["src/policykit/limits.py", "limits.py"], "kind": "file", "value": "src/policykit/limits.py"},...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["schema.py", "src/policykit/schema.py"...
ctx_compress-s1-00664
lotworks
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] restock_reorder returns the upper bound for every value instead of clamping into the range. It lives in src/lotworks/core.py; `python -m pytest tests/test_core.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the lot...
26,166
1,308
39
48
false
synth:6483676123267622608@1
synth
[{"aliases": ["src/lotworks/core.py", "core.py"], "kind": "file", "value": "src/lotworks/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["docs/changelog.md", "changelog.md"], "kind": "file", "value": "docs/CHANGELOG.md"}, {"aliases": ["src/lotw...
[{"answers": ["src/lotworks/core.py", "core.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's most re...
ctx_compress-s1-00664-twin
lotworks
ctx_compress-s1-00664
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] restock_reorder returns the upper bound for every value instead of clamping into the range. It lives in src/lotworks/core.py; `python -m pytest tests/test_core.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the lot...
15,149
1,000
27
48
false
synth:6483676123267622608@1
synth
[{"aliases": ["src/lotworks/core.py", "core.py"], "kind": "file", "value": "src/lotworks/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["docs/changelog.md", "changelog.md"], "kind": "file", "value": "docs/CHANGELOG.md"}, {"aliases": ["src/lotw...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["core.py", "src/lotworks/core.py"], "k...
ctx_compress-s1-00665
ledgerdesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] reconcile_payout_once raises ValueError when a value itself contains an equals sign. It lives in src/ledgerdesk/store.py; `python -m pytest tests/test_store.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the ledger...
68,480
3,072
110
119
false
synth:4583167613989107130@1.1649
synth
[{"aliases": ["src/ledgerdesk/store.py", "store.py"], "kind": "file", "value": "src/ledgerdesk/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["data/balance_4.csv", "balance_4.csv"], "kind": "file", "value": "data/balance_4.csv"}, {"aliases...
[{"answers": ["src/ledgerdesk/store.py", "store.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's mos...
ctx_compress-s1-00665-twin
ledgerdesk
ctx_compress-s1-00665
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] reconcile_payout_once raises ValueError when a value itself contains an equals sign. It lives in src/ledgerdesk/store.py; `python -m pytest tests/test_store.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the ledger...
31,703
1,585
64
119
false
synth:4583167613989107130@1.1649
synth
[{"aliases": ["src/ledgerdesk/store.py", "store.py"], "kind": "file", "value": "src/ledgerdesk/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["data/balance_4.csv", "balance_4.csv"], "kind": "file", "value": "data/balance_4.csv"}, {"aliases...
[{"answers": ["src/ledgerdesk/store.py", "store.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's mos...
ctx_compress-s1-00666
ledgerforge
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] reconcile_balance raises ValueError when a value itself contains an equals sign. It lives in src/ledgerforge/store.py; `python -m pytest tests/test_store.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the ledgerfor...
52,131
2,606
106
113
false
synth:5738889716072585787@1.01355
synth
[{"aliases": ["src/ledgerforge/store.py", "store.py"], "kind": "file", "value": "src/ledgerforge/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/ledgerforge/tables.py", "tables.py"], "kind": "file", "value": "src/ledgerforge/tables.py"...
[{"answers": ["src/ledgerforge/store.py", "store.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's mo...
ctx_compress-s1-00666-twin
ledgerforge
ctx_compress-s1-00666
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] reconcile_balance raises ValueError when a value itself contains an equals sign. It lives in src/ledgerforge/store.py; `python -m pytest tests/test_store.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the ledgerfor...
23,685
1,184
51
113
false
synth:5738889716072585787@1.01355
synth
[{"aliases": ["src/ledgerforge/store.py", "store.py"], "kind": "file", "value": "src/ledgerforge/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/ledgerforge/tables.py", "tables.py"], "kind": "file", "value": "src/ledgerforge/tables.py"...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["data/ledger_1.csv", "ledger_1.csv"], ...
ctx_compress-s1-00667
bucketdesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] downsample_sample produces doubled dashes for consecutive separators. It lives in src/bucketdesk/rules.py; `python -m pytest tests/test_rules.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the bucketdesk layout. [t...
84,426
3,072
108
115
false
synth:7069401691630613391@1.8108
synth
[{"aliases": ["src/bucketdesk/rules.py", "rules.py"], "kind": "file", "value": "src/bucketdesk/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/bucketdesk/io.py", "io.py"], "kind": "file", "value": "src/bucketdesk/io.py"}, {"aliases": [...
[{"answers": ["src/bucketdesk/rules.py", "rules.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's mos...
ctx_compress-s1-00667-twin
bucketdesk
ctx_compress-s1-00667
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] downsample_sample produces doubled dashes for consecutive separators. It lives in src/bucketdesk/rules.py; `python -m pytest tests/test_rules.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the bucketdesk layout. [t...
47,267
2,363
62
115
false
synth:7069401691630613391@1.8108
synth
[{"aliases": ["src/bucketdesk/rules.py", "rules.py"], "kind": "file", "value": "src/bucketdesk/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/bucketdesk/io.py", "io.py"], "kind": "file", "value": "src/bucketdesk/io.py"}, {"aliases": [...
[{"answers": ["src/bucketdesk/rules.py", "rules.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's mos...
ctx_compress-s1-00668
refundforge
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] reconcile_receipt averages one value too many once the window is full. It lives in src/refundforge/cli.py; `python -m pytest tests/test_cli.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the refundforge layout. [to...
78,240
3,072
90
97
false
synth:6786722440604216156@2.18933
synth
[{"aliases": ["src/refundforge/cli.py", "cli.py"], "kind": "file", "value": "src/refundforge/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["src/refundforge/store.py", "store.py"], "kind": "file", "value": "src/refundforge/store.py"}, {"aliases": ...
[{"answers": ["src/refundforge/cli.py", "cli.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's most r...
ctx_compress-s1-00668-twin
refundforge
ctx_compress-s1-00668
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] reconcile_receipt averages one value too many once the window is full. It lives in src/refundforge/cli.py; `python -m pytest tests/test_cli.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the refundforge layout. [to...
42,788
2,139
55
97
false
synth:6786722440604216156@2.18933
synth
[{"aliases": ["src/refundforge/cli.py", "cli.py"], "kind": "file", "value": "src/refundforge/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["src/refundforge/store.py", "store.py"], "kind": "file", "value": "src/refundforge/store.py"}, {"aliases": ...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["data/ledger_3.csv", "ledger_3.csv"], ...
ctx_compress-s1-00669
draftdesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] redact_index skips the first page; page 1 returns what should be page 2. It lives in src/draftdesk/compat.py; `python -m pytest tests/test_compat.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the draftdesk layout....
33,856
1,692
62
75
false
synth:236472732761737361@1
synth
[{"aliases": ["src/draftdesk/compat.py", "compat.py"], "kind": "file", "value": "src/draftdesk/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["src/draftdesk/audit.py", "audit.py"], "kind": "file", "value": "src/draftdesk/audit.py"}, {"...
[{"answers": ["src/draftdesk/compat.py", "compat.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's mo...
ctx_compress-s1-00669-twin
draftdesk
ctx_compress-s1-00669
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] redact_index skips the first page; page 1 returns what should be page 2. It lives in src/draftdesk/compat.py; `python -m pytest tests/test_compat.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the draftdesk layout....
13,934
1,000
36
75
false
synth:236472732761737361@1
synth
[{"aliases": ["src/draftdesk/compat.py", "compat.py"], "kind": "file", "value": "src/draftdesk/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["src/draftdesk/audit.py", "audit.py"], "kind": "file", "value": "src/draftdesk/audit.py"}, {"...
[{"answers": ["src/draftdesk/compat.py", "compat.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's mo...
ctx_compress-s1-00670
stemflow
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] wrap_snippet returns one item too few on every page. It lives in src/stemflow/cli.py; `python -m pytest tests/test_cli.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the stemflow layout. [tool_use kind=search name=...
107,557
3,072
114
143
false
synth:2602368024686319226@2.06743
synth
[{"aliases": ["src/stemflow/cli.py", "cli.py"], "kind": "file", "value": "src/stemflow/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["src/stemflow/compat.py", "compat.py"], "kind": "file", "value": "src/stemflow/compat.py"}, {"aliases": ["data/le...
[{"answers": ["src/stemflow/cli.py", "cli.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most rece...
ctx_compress-s1-00670-twin
stemflow
ctx_compress-s1-00670
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] wrap_snippet returns one item too few on every page. It lives in src/stemflow/cli.py; `python -m pytest tests/test_cli.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the stemflow layout. [tool_use kind=search name=...
75,558
3,072
84
143
false
synth:2602368024686319226@2.06743
synth
[{"aliases": ["src/stemflow/cli.py", "cli.py"], "kind": "file", "value": "src/stemflow/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["src/stemflow/compat.py", "compat.py"], "kind": "file", "value": "src/stemflow/compat.py"}, {"aliases": ["data/le...
[{"answers": ["src/stemflow/cli.py", "cli.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most rece...
ctx_compress-s1-00671
ledgerworks
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] accrue_ledger lets defaults overwrite explicit overrides. It lives in src/ledgerworks/format.py; `python -m pytest tests/test_format.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the ledgerworks layout. [tool_use ...
87,768
3,072
89
108
false
synth:7859489595642588542@1.57575
synth
[{"aliases": ["src/ledgerworks/format.py", "format.py"], "kind": "file", "value": "src/ledgerworks/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["data/refund_12.csv", "refund_12.csv"], "kind": "file", "value": "data/refund_12.csv"}, {...
[{"answers": ["src/ledgerworks/format.py", "format.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's ...
ctx_compress-s1-00671-twin
ledgerworks
ctx_compress-s1-00671
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] accrue_ledger lets defaults overwrite explicit overrides. It lives in src/ledgerworks/format.py; `python -m pytest tests/test_format.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the ledgerworks layout. [tool_use ...
50,280
2,514
55
108
false
synth:7859489595642588542@1.57575
synth
[{"aliases": ["src/ledgerworks/format.py", "format.py"], "kind": "file", "value": "src/ledgerworks/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["data/refund_12.csv", "refund_12.csv"], "kind": "file", "value": "data/refund_12.csv"}, {...
[{"answers": ["src/ledgerworks/format.py", "format.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's ...
ctx_compress-s1-00672
palletledger
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] pick_stock misses keys that differ only in case. It lives in src/palletledger/format.py; `python -m pytest tests/test_format.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the palletledger layout. [tool_use kind=se...
102,236
3,072
112
125
false
synth:8479501969019393684@2.03187
synth
[{"aliases": ["src/palletledger/format.py", "format.py"], "kind": "file", "value": "src/palletledger/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["data/stock_3.csv", "stock_3.csv"], "kind": "file", "value": "data/stock_3.csv"}, {"ali...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["changelog.md", "docs/changelog.md"], ...
ctx_compress-s1-00672-twin
palletledger
ctx_compress-s1-00672
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] pick_stock misses keys that differ only in case. It lives in src/palletledger/format.py; `python -m pytest tests/test_format.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the palletledger layout. [tool_use kind=se...
64,815
3,072
68
125
false
synth:8479501969019393684@2.03187
synth
[{"aliases": ["src/palletledger/format.py", "format.py"], "kind": "file", "value": "src/palletledger/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["data/stock_3.csv", "stock_3.csv"], "kind": "file", "value": "data/stock_3.csv"}, {"ali...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["data/warehouse_6.csv", "warehouse_6.c...
ctx_compress-s1-00673
lexemeledger
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] dedupe_glyph_page double counts carried amounts when the period closes on a weekend. It lives in src/lexemeledger/store.py; `python -m pytest tests/test_store.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the lexe...
110,439
3,072
106
121
false
synth:8452573069315062983@2.1084
synth
[{"aliases": ["src/lexemeledger/store.py", "store.py"], "kind": "file", "value": "src/lexemeledger/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/lexemeledger/format.py", "format.py"], "kind": "file", "value": "src/lexemeledger/format...
[{"answers": ["src/lexemeledger/store.py", "store.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's m...
ctx_compress-s1-00673-twin
lexemeledger
ctx_compress-s1-00673
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] dedupe_glyph_page double counts carried amounts when the period closes on a weekend. It lives in src/lexemeledger/store.py; `python -m pytest tests/test_store.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the lexe...
56,028
2,801
64
121
false
synth:8452573069315062983@2.1084
synth
[{"aliases": ["src/lexemeledger/store.py", "store.py"], "kind": "file", "value": "src/lexemeledger/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/lexemeledger/format.py", "format.py"], "kind": "file", "value": "src/lexemeledger/format...
[{"answers": ["src/lexemeledger/store.py", "store.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's m...
ctx_compress-s1-00674
stemkit
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] normalise_cue lets defaults overwrite explicit overrides. It lives in src/stemkit/format.py; `python -m pytest tests/test_format.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the stemkit layout. [tool_use kind=sea...
75,741
3,072
89
108
false
synth:3885691027885810318@1.96995
synth
[{"aliases": ["src/stemkit/format.py", "format.py"], "kind": "file", "value": "src/stemkit/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["data/snippet_4.csv", "snippet_4.csv"], "kind": "file", "value": "data/snippet_4.csv"}, {"aliases...
[{"answers": ["src/stemkit/format.py", "format.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most...
ctx_compress-s1-00674-twin
stemkit
ctx_compress-s1-00674
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] normalise_cue lets defaults overwrite explicit overrides. It lives in src/stemkit/format.py; `python -m pytest tests/test_format.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the stemkit layout. [tool_use kind=sea...
43,777
2,188
53
108
false
synth:3885691027885810318@1.96995
synth
[{"aliases": ["src/stemkit/format.py", "format.py"], "kind": "file", "value": "src/stemkit/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["data/snippet_4.csv", "snippet_4.csv"], "kind": "file", "value": "data/snippet_4.csv"}, {"aliases...
[{"answers": ["src/stemkit/format.py", "format.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most...
ctx_compress-s1-00675
marginforge
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] fold_draft averages one value too many once the window is full. It lives in src/marginforge/rules.py; `python -m pytest tests/test_rules.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the marginforge layout. [tool_...
45,516
2,275
94
103
false
synth:7422521410981315791@1
synth
[{"aliases": ["src/marginforge/rules.py", "rules.py"], "kind": "file", "value": "src/marginforge/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/marginforge/cli.py", "cli.py"], "kind": "file", "value": "src/marginforge/cli.py"}, {"alia...
[{"answers": ["src/marginforge/rules.py", "rules.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's mo...
ctx_compress-s1-00675-twin
marginforge
ctx_compress-s1-00675
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] fold_draft averages one value too many once the window is full. It lives in src/marginforge/rules.py; `python -m pytest tests/test_rules.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the marginforge layout. [tool_...
26,127
1,306
60
103
false
synth:7422521410981315791@1
synth
[{"aliases": ["src/marginforge/rules.py", "rules.py"], "kind": "file", "value": "src/marginforge/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/marginforge/cli.py", "cli.py"], "kind": "file", "value": "src/marginforge/cli.py"}, {"alia...
[{"answers": ["src/marginforge/rules.py", "rules.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's mo...
ctx_compress-s1-00676
calendarledger
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] balance_shift leaves a trailing dash when the title ends in punctuation. It lives in src/calendarledger/report.py; `python -m pytest tests/test_report.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the calendarledg...
85,817
3,072
97
106
false
synth:362992579885409264@1.57917
synth
[{"aliases": ["src/calendarledger/report.py", "report.py"], "kind": "file", "value": "src/calendarledger/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["src/calendarledger/format.py", "format.py"], "kind": "file", "value": "src/calenda...
[{"answers": ["src/calendarledger/report.py", "report.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent...
ctx_compress-s1-00676-twin
calendarledger
ctx_compress-s1-00676
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] balance_shift leaves a trailing dash when the title ends in punctuation. It lives in src/calendarledger/report.py; `python -m pytest tests/test_report.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the calendarledg...
47,178
2,358
61
106
false
synth:362992579885409264@1.57917
synth
[{"aliases": ["src/calendarledger/report.py", "report.py"], "kind": "file", "value": "src/calendarledger/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["src/calendarledger/format.py", "format.py"], "kind": "file", "value": "src/calenda...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["data/roster_1.csv", "roster_1.csv"], ...
ctx_compress-s1-00677
glyphforge
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] normalise_stem leaves a trailing dash when the title ends in punctuation. It lives in src/glyphforge/format.py; `python -m pytest tests/test_format.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the glyphforge layo...
77,559
3,072
94
117
false
synth:4992803480498594646@1.82695
synth
[{"aliases": ["src/glyphforge/format.py", "format.py"], "kind": "file", "value": "src/glyphforge/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["data/caption_6.csv", "caption_6.csv"], "kind": "file", "value": "data/caption_6.csv"}, {"a...
[{"answers": ["src/glyphforge/format.py", "format.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's m...
ctx_compress-s1-00677-twin
glyphforge
ctx_compress-s1-00677
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] normalise_stem leaves a trailing dash when the title ends in punctuation. It lives in src/glyphforge/format.py; `python -m pytest tests/test_format.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the glyphforge layo...
42,923
2,146
58
117
false
synth:4992803480498594646@1.82695
synth
[{"aliases": ["src/glyphforge/format.py", "format.py"], "kind": "file", "value": "src/glyphforge/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["data/caption_6.csv", "caption_6.csv"], "kind": "file", "value": "data/caption_6.csv"}, {"a...
[{"answers": ["src/glyphforge/format.py", "format.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's m...
ctx_compress-s1-00678
slugdesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] hyphenate_chunk returns the upper bound for every value instead of clamping into the range. It lives in src/slugdesk/compat.py; `python -m pytest tests/test_compat.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the...
67,074
3,072
96
111
false
synth:8568003422698135260@1.32927
synth
[{"aliases": ["src/slugdesk/compat.py", "compat.py"], "kind": "file", "value": "src/slugdesk/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["src/slugdesk/cli.py", "cli.py"], "kind": "file", "value": "src/slugdesk/cli.py"}, {"aliases": ...
[{"answers": ["src/slugdesk/compat.py", "compat.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's mos...
ctx_compress-s1-00678-twin
slugdesk
ctx_compress-s1-00678
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] hyphenate_chunk returns the upper bound for every value instead of clamping into the range. It lives in src/slugdesk/compat.py; `python -m pytest tests/test_compat.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the...
36,439
1,821
60
111
false
synth:8568003422698135260@1.32927
synth
[{"aliases": ["src/slugdesk/compat.py", "compat.py"], "kind": "file", "value": "src/slugdesk/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["src/slugdesk/cli.py", "cli.py"], "kind": "file", "value": "src/slugdesk/cli.py"}, {"aliases": ...
[{"answers": ["src/slugdesk/compat.py", "compat.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's mos...
ctx_compress-s1-00679
corridorworks
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] snap_tile skips the first page; page 1 returns what should be page 2. It lives in src/corridorworks/core.py; `python -m pytest tests/test_core.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the corridorworks layout...
38,986
1,949
65
70
false
synth:7156655746910016346@1
synth
[{"aliases": ["src/corridorworks/core.py", "core.py"], "kind": "file", "value": "src/corridorworks/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/corridorworks/format.py", "format.py"], "kind": "file", "value": "src/corridorworks/format.p...
[{"answers": ["src/corridorworks/core.py", "core.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's mo...
ctx_compress-s1-00679-twin
corridorworks
ctx_compress-s1-00679
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] snap_tile skips the first page; page 1 returns what should be page 2. It lives in src/corridorworks/core.py; `python -m pytest tests/test_core.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the corridorworks layout...
24,363
1,218
41
70
false
synth:7156655746910016346@1
synth
[{"aliases": ["src/corridorworks/core.py", "core.py"], "kind": "file", "value": "src/corridorworks/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/corridorworks/format.py", "format.py"], "kind": "file", "value": "src/corridorworks/format.p...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["schema.py", "src/corridorworks/schema...
ctx_compress-s1-00680
policyworks
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] revoke_session produces doubled dashes for consecutive separators. It lives in src/policyworks/rules.py; `python -m pytest tests/test_rules.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the policyworks layout. [to...
47,809
2,390
80
101
false
synth:6460049488044361260@1.09238
synth
[{"aliases": ["src/policyworks/rules.py", "rules.py"], "kind": "file", "value": "src/policyworks/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/policyworks/io.py", "io.py"], "kind": "file", "value": "src/policyworks/io.py"}, {"aliases...
[{"answers": ["src/policyworks/rules.py", "rules.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's mo...
ctx_compress-s1-00680-twin
policyworks
ctx_compress-s1-00680
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] revoke_session produces doubled dashes for consecutive separators. It lives in src/policyworks/rules.py; `python -m pytest tests/test_rules.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the policyworks layout. [to...
28,306
1,415
46
101
false
synth:6460049488044361260@1.09238
synth
[{"aliases": ["src/policyworks/rules.py", "rules.py"], "kind": "file", "value": "src/policyworks/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/policyworks/io.py", "io.py"], "kind": "file", "value": "src/policyworks/io.py"}, {"aliases...
[{"answers": ["src/policyworks/rules.py", "rules.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's mo...
ctx_compress-s1-00681
policyflow
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] rotate_session returns one item too few on every page. It lives in src/policyflow/compat.py; `python -m pytest tests/test_compat.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the policyflow layout. [tool_use kind=...
116,480
3,072
142
155
false
synth:7737103266175032013@2.1175
synth
[{"aliases": ["src/policyflow/compat.py", "compat.py"], "kind": "file", "value": "src/policyflow/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["data/nonce_11.csv", "nonce_11.csv"], "kind": "file", "value": "data/nonce_11.csv"}, {"alia...
[{"answers": ["src/policyflow/compat.py", "compat.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's m...
ctx_compress-s1-00681-twin
policyflow
ctx_compress-s1-00681
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] rotate_session returns one item too few on every page. It lives in src/policyflow/compat.py; `python -m pytest tests/test_compat.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the policyflow layout. [tool_use kind=...
59,075
2,953
74
155
false
synth:7737103266175032013@2.1175
synth
[{"aliases": ["src/policyflow/compat.py", "compat.py"], "kind": "file", "value": "src/policyflow/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["data/nonce_11.csv", "nonce_11.csv"], "kind": "file", "value": "data/nonce_11.csv"}, {"alia...
[{"answers": ["src/policyflow/compat.py", "compat.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's m...
ctx_compress-s1-00682
quotaflow
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] rotate_slot raises KeyError for a missing key instead of returning the default. It lives in src/quotaflow/report.py; `python -m pytest tests/test_report.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the quotaflow ...
105,977
3,072
120
137
false
synth:5611699370435956919@2.2446
synth
[{"aliases": ["src/quotaflow/report.py", "report.py"], "kind": "file", "value": "src/quotaflow/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["data/window_3.csv", "window_3.csv"], "kind": "file", "value": "data/window_3.csv"}, {"aliase...
[{"answers": ["src/quotaflow/report.py", "report.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's mo...
ctx_compress-s1-00682-twin
quotaflow
ctx_compress-s1-00682
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] rotate_slot raises KeyError for a missing key instead of returning the default. It lives in src/quotaflow/report.py; `python -m pytest tests/test_report.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the quotaflow ...
62,985
3,072
80
137
false
synth:5611699370435956919@2.2446
synth
[{"aliases": ["src/quotaflow/report.py", "report.py"], "kind": "file", "value": "src/quotaflow/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["data/window_3.csv", "window_3.csv"], "kind": "file", "value": "data/window_3.csv"}, {"aliase...
[{"answers": ["src/quotaflow/report.py", "report.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's mo...
ctx_compress-s1-00683
clauseforge
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] paginate_section_window leaves a trailing dash when the title ends in punctuation. It lives in src/clauseforge/limits.py; `python -m pytest tests/test_limits.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the claus...
33,403
1,670
68
85
false
synth:460765996291846692@1
synth
[{"aliases": ["src/clauseforge/limits.py", "limits.py"], "kind": "file", "value": "src/clauseforge/limits.py"}, {"aliases": ["tests/test_limits.py", "test_limits.py"], "kind": "file", "value": "tests/test_limits.py"}, {"aliases": ["src/clauseforge/core.py", "core.py"], "kind": "file", "value": "src/clauseforge/core.py"...
[{"answers": ["src/clauseforge/limits.py", "limits.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's ...
ctx_compress-s1-00683-twin
clauseforge
ctx_compress-s1-00683
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] paginate_section_window leaves a trailing dash when the title ends in punctuation. It lives in src/clauseforge/limits.py; `python -m pytest tests/test_limits.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the claus...
25,188
1,259
46
85
false
synth:460765996291846692@1
synth
[{"aliases": ["src/clauseforge/limits.py", "limits.py"], "kind": "file", "value": "src/clauseforge/limits.py"}, {"aliases": ["tests/test_limits.py", "test_limits.py"], "kind": "file", "value": "tests/test_limits.py"}, {"aliases": ["src/clauseforge/core.py", "core.py"], "kind": "file", "value": "src/clauseforge/core.py"...
[{"answers": ["src/clauseforge/limits.py", "limits.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's ...
ctx_compress-s1-00684
stemledger
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] truncate_chunk returns one item too few on every page. It lives in src/stemledger/compat.py; `python -m pytest tests/test_compat.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the stemledger layout. [tool_use kind=...
99,056
3,072
117
148
false
synth:469101768560613821@2.17457
synth
[{"aliases": ["src/stemledger/compat.py", "compat.py"], "kind": "file", "value": "src/stemledger/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["src/stemledger/schema.py", "schema.py"], "kind": "file", "value": "src/stemledger/schema.p...
[{"answers": ["src/stemledger/compat.py", "compat.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's m...
ctx_compress-s1-00684-twin
stemledger
ctx_compress-s1-00684
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] truncate_chunk returns one item too few on every page. It lives in src/stemledger/compat.py; `python -m pytest tests/test_compat.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the stemledger layout. [tool_use kind=...
59,449
2,972
79
148
false
synth:469101768560613821@2.17457
synth
[{"aliases": ["src/stemledger/compat.py", "compat.py"], "kind": "file", "value": "src/stemledger/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["src/stemledger/schema.py", "schema.py"], "kind": "file", "value": "src/stemledger/schema.p...
[{"answers": ["src/stemledger/compat.py", "compat.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's m...
ctx_compress-s1-00685
bookingflow
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] balance_shift raises ZeroDivisionError on the first value and lags by one afterwards. It lives in src/bookingflow/io.py; `python -m pytest tests/test_io.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the bookingflo...
72,481
3,072
123
134
false
synth:4815165967927037952@1.8762
synth
[{"aliases": ["src/bookingflow/io.py", "io.py"], "kind": "file", "value": "src/bookingflow/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["src/bookingflow/compat.py", "compat.py"], "kind": "file", "value": "src/bookingflow/compat.py"}, {"aliases": ["s...
[{"answers": ["src/bookingflow/io.py", "io.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most rec...
ctx_compress-s1-00685-twin
bookingflow
ctx_compress-s1-00685
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] balance_shift raises ZeroDivisionError on the first value and lags by one afterwards. It lives in src/bookingflow/io.py; `python -m pytest tests/test_io.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the bookingflo...
37,651
1,882
73
134
false
synth:4815165967927037952@1.8762
synth
[{"aliases": ["src/bookingflow/io.py", "io.py"], "kind": "file", "value": "src/bookingflow/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["src/bookingflow/compat.py", "compat.py"], "kind": "file", "value": "src/bookingflow/compat.py"}, {"aliases": ["s...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["test_audit.py", "tests/test_audit.py"...
ctx_compress-s1-00686
beaconledger
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] interpolate_corridor_once leaves a trailing dash when the title ends in punctuation. It lives in src/beaconledger/format.py; `python -m pytest tests/test_format.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the be...
101,754
3,072
114
135
false
synth:2696668864717970557@2.1953
synth
[{"aliases": ["src/beaconledger/format.py", "format.py"], "kind": "file", "value": "src/beaconledger/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["data/corridor_0.csv", "corridor_0.csv"], "kind": "file", "value": "data/corridor_0.csv...
[{"answers": ["src/beaconledger/format.py", "format.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's...
ctx_compress-s1-00686-twin
beaconledger
ctx_compress-s1-00686
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] interpolate_corridor_once leaves a trailing dash when the title ends in punctuation. It lives in src/beaconledger/format.py; `python -m pytest tests/test_format.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the be...
56,002
2,800
72
135
false
synth:2696668864717970557@2.1953
synth
[{"aliases": ["src/beaconledger/format.py", "format.py"], "kind": "file", "value": "src/beaconledger/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["data/corridor_0.csv", "corridor_0.csv"], "kind": "file", "value": "data/corridor_0.csv...
[{"answers": ["src/beaconledger/format.py", "format.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's...
ctx_compress-s1-00687
receiptkit
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] allocate_credit ignores the cap, so late attempts wait for minutes. It lives in src/receiptkit/report.py; `python -m pytest tests/test_report.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the receiptkit layout. [t...
93,788
3,072
107
120
false
synth:33305216687056093@1.58007
synth
[{"aliases": ["src/receiptkit/report.py", "report.py"], "kind": "file", "value": "src/receiptkit/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["src/receiptkit/schema.py", "schema.py"], "kind": "file", "value": "src/receiptkit/schema.p...
[{"answers": ["src/receiptkit/report.py", "report.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's m...
ctx_compress-s1-00687-twin
receiptkit
ctx_compress-s1-00687
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] allocate_credit ignores the cap, so late attempts wait for minutes. It lives in src/receiptkit/report.py; `python -m pytest tests/test_report.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the receiptkit layout. [t...
51,026
2,551
67
120
false
synth:33305216687056093@1.58007
synth
[{"aliases": ["src/receiptkit/report.py", "report.py"], "kind": "file", "value": "src/receiptkit/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["src/receiptkit/schema.py", "schema.py"], "kind": "file", "value": "src/receiptkit/schema.p...
[{"answers": ["src/receiptkit/report.py", "report.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's m...
ctx_compress-s1-00688
ledgerforge
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] roll_receipt raises ZeroDivisionError on the first value and lags by one afterwards. It lives in src/ledgerforge/report.py; `python -m pytest tests/test_report.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the led...
38,575
1,928
64
71
false
synth:5066266228216269295@1
synth
[{"aliases": ["src/ledgerforge/report.py", "report.py"], "kind": "file", "value": "src/ledgerforge/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["src/ledgerforge/io.py", "io.py"], "kind": "file", "value": "src/ledgerforge/io.py"}, {"a...
[{"answers": ["src/ledgerforge/report.py", "report.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's ...
ctx_compress-s1-00688-twin
ledgerforge
ctx_compress-s1-00688
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] roll_receipt raises ZeroDivisionError on the first value and lags by one afterwards. It lives in src/ledgerforge/report.py; `python -m pytest tests/test_report.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the led...
22,730
1,136
38
71
false
synth:5066266228216269295@1
synth
[{"aliases": ["src/ledgerforge/report.py", "report.py"], "kind": "file", "value": "src/ledgerforge/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["src/ledgerforge/io.py", "io.py"], "kind": "file", "value": "src/ledgerforge/io.py"}, {"a...
[{"answers": ["src/ledgerforge/report.py", "report.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's ...
ctx_compress-s1-00689
gaugeworks
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] downsample_series double counts carried amounts when the period closes on a weekend. It lives in src/gaugeworks/core.py; `python -m pytest tests/test_core.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the gaugewor...
88,854
3,072
102
129
false
synth:7581956339600745482@1.62542
synth
[{"aliases": ["src/gaugeworks/core.py", "core.py"], "kind": "file", "value": "src/gaugeworks/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/gaugeworks/limits.py", "limits.py"], "kind": "file", "value": "src/gaugeworks/limits.py"}, {"alias...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["data/trace_13.csv", "trace_13.csv"], ...
ctx_compress-s1-00689-twin
gaugeworks
ctx_compress-s1-00689
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] downsample_series double counts carried amounts when the period closes on a weekend. It lives in src/gaugeworks/core.py; `python -m pytest tests/test_core.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the gaugewor...
50,667
2,533
62
129
false
synth:7581956339600745482@1.62542
synth
[{"aliases": ["src/gaugeworks/core.py", "core.py"], "kind": "file", "value": "src/gaugeworks/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/gaugeworks/limits.py", "limits.py"], "kind": "file", "value": "src/gaugeworks/limits.py"}, {"alias...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["changelog.md", "docs/changelog.md"], ...
ctx_compress-s1-00690
tariffledger
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] allocate_tariff lets defaults overwrite explicit overrides. It lives in src/tariffledger/rules.py; `python -m pytest tests/test_rules.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the tariffledger layout. [tool_us...
40,909
2,045
61
68
false
synth:3595023647315032647@1
synth
[{"aliases": ["src/tariffledger/rules.py", "rules.py"], "kind": "file", "value": "src/tariffledger/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/tariffledger/schema.py", "schema.py"], "kind": "file", "value": "src/tariffledger/schema...
[{"answers": ["src/tariffledger/rules.py", "rules.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's m...
ctx_compress-s1-00690-twin
tariffledger
ctx_compress-s1-00690
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] allocate_tariff lets defaults overwrite explicit overrides. It lives in src/tariffledger/rules.py; `python -m pytest tests/test_rules.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the tariffledger layout. [tool_us...
14,643
1,000
31
68
false
synth:3595023647315032647@1
synth
[{"aliases": ["src/tariffledger/rules.py", "rules.py"], "kind": "file", "value": "src/tariffledger/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/tariffledger/schema.py", "schema.py"], "kind": "file", "value": "src/tariffledger/schema...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["data/tariff_1.csv", "tariff_1.csv"], ...
ctx_compress-s1-00691
slugledger
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] normalise_cue returns one item too few on every page. It lives in src/slugledger/cli.py; `python -m pytest tests/test_cli.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the slugledger layout. [tool_use kind=search ...
112,555
3,072
140
153
false
synth:6085195623728166495@1.92732
synth
[{"aliases": ["src/slugledger/cli.py", "cli.py"], "kind": "file", "value": "src/slugledger/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["src/slugledger/rules.py", "rules.py"], "kind": "file", "value": "src/slugledger/rules.py"}, {"aliases": ["sr...
[{"answers": ["src/slugledger/cli.py", "cli.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most re...
ctx_compress-s1-00691-twin
slugledger
ctx_compress-s1-00691
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] normalise_cue returns one item too few on every page. It lives in src/slugledger/cli.py; `python -m pytest tests/test_cli.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the slugledger layout. [tool_use kind=search ...
54,620
2,731
70
153
false
synth:6085195623728166495@1.92732
synth
[{"aliases": ["src/slugledger/cli.py", "cli.py"], "kind": "file", "value": "src/slugledger/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["src/slugledger/rules.py", "rules.py"], "kind": "file", "value": "src/slugledger/rules.py"}, {"aliases": ["sr...
[{"answers": ["src/slugledger/cli.py", "cli.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's most re...
ctx_compress-s1-00692
seriesdesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] rate_gauge raises KeyError for a missing key instead of returning the default. It lives in src/seriesdesk/rules.py; `python -m pytest tests/test_rules.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the seriesdesk l...
85,227
3,072
109
138
false
synth:8796758942484349903@1.97065
synth
[{"aliases": ["src/seriesdesk/rules.py", "rules.py"], "kind": "file", "value": "src/seriesdesk/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["data/probe_6.csv", "probe_6.csv"], "kind": "file", "value": "data/probe_6.csv"}, {"aliases": ["d...
[{"answers": ["src/seriesdesk/rules.py", "rules.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's mos...
ctx_compress-s1-00692-twin
seriesdesk
ctx_compress-s1-00692
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] rate_gauge raises KeyError for a missing key instead of returning the default. It lives in src/seriesdesk/rules.py; `python -m pytest tests/test_rules.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the seriesdesk l...
44,656
2,232
65
138
false
synth:8796758942484349903@1.97065
synth
[{"aliases": ["src/seriesdesk/rules.py", "rules.py"], "kind": "file", "value": "src/seriesdesk/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["data/probe_6.csv", "probe_6.csv"], "kind": "file", "value": "data/probe_6.csv"}, {"aliases": ["d...
[{"answers": ["src/seriesdesk/rules.py", "rules.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's mos...
ctx_compress-s1-00693
rosterledger
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] assign_calendar_window skips the first page; page 1 returns what should be page 2. It lives in src/rosterledger/compat.py; `python -m pytest tests/test_compat.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the rost...
69,700
3,072
100
121
false
synth:3900506804169254237@1.52872
synth
[{"aliases": ["src/rosterledger/compat.py", "compat.py"], "kind": "file", "value": "src/rosterledger/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["src/rosterledger/schema.py", "schema.py"], "kind": "file", "value": "src/rosterledger/...
[{"answers": ["src/rosterledger/compat.py", "compat.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's...
ctx_compress-s1-00693-twin
rosterledger
ctx_compress-s1-00693
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] assign_calendar_window skips the first page; page 1 returns what should be page 2. It lives in src/rosterledger/compat.py; `python -m pytest tests/test_compat.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the rost...
46,222
2,311
72
121
false
synth:3900506804169254237@1.52872
synth
[{"aliases": ["src/rosterledger/compat.py", "compat.py"], "kind": "file", "value": "src/rosterledger/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["src/rosterledger/schema.py", "schema.py"], "kind": "file", "value": "src/rosterledger/...
[{"answers": ["src/rosterledger/compat.py", "compat.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's...
ctx_compress-s1-00694
shiftledger
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] rotate_cadence raises KeyError for a missing key instead of returning the default. It lives in src/shiftledger/compat.py; `python -m pytest tests/test_compat.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the shift...
64,091
3,072
92
103
false
synth:7914578249798907969@1.02372
synth
[{"aliases": ["src/shiftledger/compat.py", "compat.py"], "kind": "file", "value": "src/shiftledger/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["data/booking_3.csv", "booking_3.csv"], "kind": "file", "value": "data/booking_3.csv"}, {...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["compat.py", "src/shiftledger/compat.p...
ctx_compress-s1-00694-twin
shiftledger
ctx_compress-s1-00694
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] rotate_cadence raises KeyError for a missing key instead of returning the default. It lives in src/shiftledger/compat.py; `python -m pytest tests/test_compat.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the shift...
38,465
1,923
56
103
false
synth:7914578249798907969@1.02372
synth
[{"aliases": ["src/shiftledger/compat.py", "compat.py"], "kind": "file", "value": "src/shiftledger/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["data/booking_3.csv", "booking_3.csv"], "kind": "file", "value": "data/booking_3.csv"}, {...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["test_cli.py", "tests/test_cli.py"], "...
ctx_compress-s1-00695
policyworks
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] revoke_nonce_once returns the upper bound for every value instead of clamping into the range. It lives in src/policyworks/core.py; `python -m pytest tests/test_core.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping th...
69,595
3,072
97
112
false
synth:7994454209633912667@2.3706
synth
[{"aliases": ["src/policyworks/core.py", "core.py"], "kind": "file", "value": "src/policyworks/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/policyworks/rules.py", "rules.py"], "kind": "file", "value": "src/policyworks/rules.py"}, {"alia...
[{"answers": ["src/policyworks/core.py", "core.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's most...
ctx_compress-s1-00695-twin
policyworks
ctx_compress-s1-00695
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] revoke_nonce_once returns the upper bound for every value instead of clamping into the range. It lives in src/policyworks/core.py; `python -m pytest tests/test_core.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping th...
43,052
2,152
61
112
false
synth:7994454209633912667@2.3706
synth
[{"aliases": ["src/policyworks/core.py", "core.py"], "kind": "file", "value": "src/policyworks/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/policyworks/rules.py", "rules.py"], "kind": "file", "value": "src/policyworks/rules.py"}, {"alia...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["data/nonce_4.csv", "nonce_4.csv"], "k...
ctx_compress-s1-00696
warehouseledger
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] count_shipment raises ZeroDivisionError on the first value and lags by one afterwards. It lives in src/warehouseledger/compat.py; `python -m pytest tests/test_compat.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping t...
96,919
3,072
132
149
false
synth:3625653088337662802@2.0958
synth
[{"aliases": ["src/warehouseledger/compat.py", "compat.py"], "kind": "file", "value": "src/warehouseledger/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["data/reorder_7.csv", "reorder_7.csv"], "kind": "file", "value": "data/reorder_7....
[{"answers": ["src/warehouseledger/compat.py", "compat.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agen...
ctx_compress-s1-00696-twin
warehouseledger
ctx_compress-s1-00696
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] count_shipment raises ZeroDivisionError on the first value and lags by one afterwards. It lives in src/warehouseledger/compat.py; `python -m pytest tests/test_compat.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping t...
54,676
2,733
76
149
false
synth:3625653088337662802@2.0958
synth
[{"aliases": ["src/warehouseledger/compat.py", "compat.py"], "kind": "file", "value": "src/warehouseledger/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["data/reorder_7.csv", "reorder_7.csv"], "kind": "file", "value": "data/reorder_7....
[{"answers": ["src/warehouseledger/compat.py", "compat.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agen...
ctx_compress-s1-00697
claimdesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] extend_nonce double counts carried amounts when the period closes on a weekend. It lives in src/claimdesk/audit.py; `python -m pytest tests/test_audit.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the claimdesk la...
59,668
2,983
92
109
false
synth:8304056822146807492@1.22505
synth
[{"aliases": ["src/claimdesk/audit.py", "audit.py"], "kind": "file", "value": "src/claimdesk/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["src/claimdesk/cli.py", "cli.py"], "kind": "file", "value": "src/claimdesk/cli.py"}, {"aliases": ["...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["data/grant_2.csv", "grant_2.csv"], "k...
ctx_compress-s1-00697-twin
claimdesk
ctx_compress-s1-00697
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] extend_nonce double counts carried amounts when the period closes on a weekend. It lives in src/claimdesk/audit.py; `python -m pytest tests/test_audit.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the claimdesk la...
31,993
1,599
52
109
false
synth:8304056822146807492@1.22505
synth
[{"aliases": ["src/claimdesk/audit.py", "audit.py"], "kind": "file", "value": "src/claimdesk/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["src/claimdesk/cli.py", "cli.py"], "kind": "file", "value": "src/claimdesk/cli.py"}, {"aliases": ["...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["test_compat.py", "tests/test_compat.p...
ctx_compress-s1-00698
skuflow
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] transfer_reorder_batch raises ZeroDivisionError on the first value and lags by one afterwards. It lives in src/skuflow/io.py; `python -m pytest tests/test_io.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the skufl...
28,066
1,403
64
71
false
synth:1615540315082104988@1
synth
[{"aliases": ["src/skuflow/io.py", "io.py"], "kind": "file", "value": "src/skuflow/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["data/sku_0.csv", "sku_0.csv"], "kind": "file", "value": "data/sku_0.csv"}, {"aliases": ["docs/changelog.md", "changelog....
[{"answers": ["src/skuflow/io.py", "io.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's most recent ...
ctx_compress-s1-00698-twin
skuflow
ctx_compress-s1-00698
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] transfer_reorder_batch raises ZeroDivisionError on the first value and lags by one afterwards. It lives in src/skuflow/io.py; `python -m pytest tests/test_io.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the skufl...
12,269
1,000
37
71
false
synth:1615540315082104988@1
synth
[{"aliases": ["src/skuflow/io.py", "io.py"], "kind": "file", "value": "src/skuflow/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["data/sku_0.csv", "sku_0.csv"], "kind": "file", "value": "data/sku_0.csv"}, {"aliases": ["docs/changelog.md", "changelog....
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["io.py", "src/skuflow/io.py"], "kind":...
ctx_compress-s1-00699
grantkit
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] extend_policy double counts carried amounts when the period closes on a weekend. It lives in src/grantkit/cli.py; `python -m pytest tests/test_cli.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the grantkit layout....
56,855
2,842
95
116
false
synth:6237819465288236767@1.05858
synth
[{"aliases": ["src/grantkit/cli.py", "cli.py"], "kind": "file", "value": "src/grantkit/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["data/grant_3.csv", "grant_3.csv"], "kind": "file", "value": "data/grant_3.csv"}, {"aliases": ["src/grantkit/audi...
[{"answers": ["src/grantkit/cli.py", "cli.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most rece...
ctx_compress-s1-00699-twin
grantkit
ctx_compress-s1-00699
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] extend_policy double counts carried amounts when the period closes on a weekend. It lives in src/grantkit/cli.py; `python -m pytest tests/test_cli.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the grantkit layout....
28,673
1,433
55
116
false
synth:6237819465288236767@1.05858
synth
[{"aliases": ["src/grantkit/cli.py", "cli.py"], "kind": "file", "value": "src/grantkit/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["data/grant_3.csv", "grant_3.csv"], "kind": "file", "value": "data/grant_3.csv"}, {"aliases": ["src/grantkit/audi...
[{"answers": ["src/grantkit/cli.py", "cli.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's most rece...
ctx_compress-s1-00700
quotaforge
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] assign_shift misses keys that differ only in case. It lives in src/quotaforge/core.py; `python -m pytest tests/test_core.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the quotaforge layout. [tool_use kind=search n...
92,449
3,072
141
150
false
synth:2465052383126070524@2.0014
synth
[{"aliases": ["src/quotaforge/core.py", "core.py"], "kind": "file", "value": "src/quotaforge/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/quotaforge/rules.py", "rules.py"], "kind": "file", "value": "src/quotaforge/rules.py"}, {"aliases"...
[{"answers": ["src/quotaforge/core.py", "core.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's most ...
ctx_compress-s1-00700-twin
quotaforge
ctx_compress-s1-00700
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] assign_shift misses keys that differ only in case. It lives in src/quotaforge/core.py; `python -m pytest tests/test_core.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the quotaforge layout. [tool_use kind=search n...
40,479
2,023
69
150
false
synth:2465052383126070524@2.0014
synth
[{"aliases": ["src/quotaforge/core.py", "core.py"], "kind": "file", "value": "src/quotaforge/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/quotaforge/rules.py", "rules.py"], "kind": "file", "value": "src/quotaforge/rules.py"}, {"aliases"...
[{"answers": ["src/quotaforge/core.py", "core.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's most ...
ctx_compress-s1-00701
tokenflow
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] extend_nonce_window leaves a trailing dash when the title ends in punctuation. It lives in src/tokenflow/core.py; `python -m pytest tests/test_core.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the tokenflow layou...
66,652
3,072
90
101
false
synth:5744747558274464866@1.02
synth
[{"aliases": ["src/tokenflow/core.py", "core.py"], "kind": "file", "value": "src/tokenflow/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/tokenflow/compat.py", "compat.py"], "kind": "file", "value": "src/tokenflow/compat.py"}, {"aliases":...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["core.py", "src/tokenflow/core.py"], "...
ctx_compress-s1-00701-twin
tokenflow
ctx_compress-s1-00701
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] extend_nonce_window leaves a trailing dash when the title ends in punctuation. It lives in src/tokenflow/core.py; `python -m pytest tests/test_core.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the tokenflow layou...
42,848
2,142
58
101
false
synth:5744747558274464866@1.02
synth
[{"aliases": ["src/tokenflow/core.py", "core.py"], "kind": "file", "value": "src/tokenflow/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/tokenflow/compat.py", "compat.py"], "kind": "file", "value": "src/tokenflow/compat.py"}, {"aliases":...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["limits.py", "src/tokenflow/limits.py"...
ctx_compress-s1-00702
grantledger
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] verify_claim returns the upper bound for every value instead of clamping into the range. It lives in src/grantledger/core.py; `python -m pytest tests/test_core.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the gra...
46,963
2,348
85
96
false
synth:6049826192142773811@1.36183
synth
[{"aliases": ["src/grantledger/core.py", "core.py"], "kind": "file", "value": "src/grantledger/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/grantledger/tables.py", "tables.py"], "kind": "file", "value": "src/grantledger/tables.py"}, {"a...
[{"answers": ["src/grantledger/core.py", "core.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's most...
ctx_compress-s1-00702-twin
grantledger
ctx_compress-s1-00702
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] verify_claim returns the upper bound for every value instead of clamping into the range. It lives in src/grantledger/core.py; `python -m pytest tests/test_core.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the gra...
33,086
1,654
49
96
false
synth:6049826192142773811@1.36183
synth
[{"aliases": ["src/grantledger/core.py", "core.py"], "kind": "file", "value": "src/grantledger/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/grantledger/tables.py", "tables.py"], "kind": "file", "value": "src/grantledger/tables.py"}, {"a...
[{"answers": ["src/grantledger/core.py", "core.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's most...
ctx_compress-s1-00703
revisionkit
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] paginate_revision raises ValueError when a value itself contains an equals sign. It lives in src/revisionkit/report.py; `python -m pytest tests/test_report.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the revisio...
111,577
3,072
139
152
false
synth:5789376268618102778@2.2148
synth
[{"aliases": ["src/revisionkit/report.py", "report.py"], "kind": "file", "value": "src/revisionkit/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["data/revision_2.csv", "revision_2.csv"], "kind": "file", "value": "data/revision_2.csv"}...
[{"answers": ["src/revisionkit/report.py", "report.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's ...
ctx_compress-s1-00703-twin
revisionkit
ctx_compress-s1-00703
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] paginate_revision raises ValueError when a value itself contains an equals sign. It lives in src/revisionkit/report.py; `python -m pytest tests/test_report.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the revisio...
69,740
3,072
89
152
false
synth:5789376268618102778@2.2148
synth
[{"aliases": ["src/revisionkit/report.py", "report.py"], "kind": "file", "value": "src/revisionkit/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["data/revision_2.csv", "revision_2.csv"], "kind": "file", "value": "data/revision_2.csv"}...
[{"answers": ["src/revisionkit/report.py", "report.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's ...
ctx_compress-s1-00704
invoicedesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] accrue_credit raises ZeroDivisionError on the first value and lags by one afterwards. It lives in src/invoicedesk/format.py; `python -m pytest tests/test_format.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the in...
67,726
3,072
83
106
false
synth:3499844801702486679@1.35725
synth
[{"aliases": ["src/invoicedesk/format.py", "format.py"], "kind": "file", "value": "src/invoicedesk/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["src/invoicedesk/cli.py", "cli.py"], "kind": "file", "value": "src/invoicedesk/cli.py"}, ...
[{"answers": ["src/invoicedesk/format.py", "format.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's ...
ctx_compress-s1-00704-twin
invoicedesk
ctx_compress-s1-00704
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] accrue_credit raises ZeroDivisionError on the first value and lags by one afterwards. It lives in src/invoicedesk/format.py; `python -m pytest tests/test_format.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the in...
37,007
1,850
61
106
false
synth:3499844801702486679@1.35725
synth
[{"aliases": ["src/invoicedesk/format.py", "format.py"], "kind": "file", "value": "src/invoicedesk/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["src/invoicedesk/cli.py", "cli.py"], "kind": "file", "value": "src/invoicedesk/cli.py"}, ...
[{"answers": ["src/invoicedesk/format.py", "format.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's ...