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-s1-00805
palletflow
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 keeps leading spaces in keys so lookups after a semicolon fail. It lives in src/palletflow/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 palletflow...
62,800
3,072
101
110
false
synth:6653378940971106951@1.33003
synth
[{"aliases": ["src/palletflow/rules.py", "rules.py"], "kind": "file", "value": "src/palletflow/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/palletflow/schema.py", "schema.py"], "kind": "file", "value": "src/palletflow/schema.py"}, {...
[{"answers": ["src/palletflow/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-s1-00805-twin
palletflow
ctx_compress-s1-00805
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 keeps leading spaces in keys so lookups after a semicolon fail. It lives in src/palletflow/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 palletflow...
32,712
1,635
59
110
false
synth:6653378940971106951@1.33003
synth
[{"aliases": ["src/palletflow/rules.py", "rules.py"], "kind": "file", "value": "src/palletflow/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/palletflow/schema.py", "schema.py"], "kind": "file", "value": "src/palletflow/schema.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": ["io.py", "src/palletflow/io.py"], "kin...
ctx_compress-s1-00806
beaconworks
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_fix leaves a trailing dash when the title ends in punctuation. It lives in src/beaconworks/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 beaconworks l...
43,131
2,156
92
113
false
synth:2899675373134466544@1.47218
synth
[{"aliases": ["src/beaconworks/limits.py", "limits.py"], "kind": "file", "value": "src/beaconworks/limits.py"}, {"aliases": ["tests/test_limits.py", "test_limits.py"], "kind": "file", "value": "tests/test_limits.py"}, {"aliases": ["data/zone_2.csv", "zone_2.csv"], "kind": "file", "value": "data/zone_2.csv"}, {"aliases"...
[{"answers": ["src/beaconworks/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-s1-00806-twin
beaconworks
ctx_compress-s1-00806
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_fix leaves a trailing dash when the title ends in punctuation. It lives in src/beaconworks/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 beaconworks l...
29,508
1,475
59
113
false
synth:2899675373134466544@1.47218
synth
[{"aliases": ["src/beaconworks/limits.py", "limits.py"], "kind": "file", "value": "src/beaconworks/limits.py"}, {"aliases": ["tests/test_limits.py", "test_limits.py"], "kind": "file", "value": "tests/test_limits.py"}, {"aliases": ["data/zone_2.csv", "zone_2.csv"], "kind": "file", "value": "data/zone_2.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/beaconworks/cli.py"], "...
ctx_compress-s1-00807
cadencekit
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_calendar raises KeyError for a missing key instead of returning the default. It lives in src/cadencekit/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 cadence...
21,985
1,099
50
61
false
synth:2675072283664341766@1
synth
[{"aliases": ["src/cadencekit/store.py", "store.py"], "kind": "file", "value": "src/cadencekit/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/cadencekit/format.py", "format.py"], "kind": "file", "value": "src/cadencekit/format.py"}, {...
[{"answers": ["src/cadencekit/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-s1-00807-twin
cadencekit
ctx_compress-s1-00807
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_calendar raises KeyError for a missing key instead of returning the default. It lives in src/cadencekit/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 cadence...
13,320
1,000
36
61
false
synth:2675072283664341766@1
synth
[{"aliases": ["src/cadencekit/store.py", "store.py"], "kind": "file", "value": "src/cadencekit/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/cadencekit/format.py", "format.py"], "kind": "file", "value": "src/cadencekit/format.py"}, {...
[{"answers": ["src/cadencekit/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-s1-00808
chunkflow
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_lexeme raises ValueError when a value itself contains an equals sign. It lives in src/chunkflow/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 chunkflow l...
43,813
2,190
68
87
false
synth:74515841235785874@1
synth
[{"aliases": ["src/chunkflow/format.py", "format.py"], "kind": "file", "value": "src/chunkflow/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["src/chunkflow/compat.py", "compat.py"], "kind": "file", "value": "src/chunkflow/compat.py"},...
[{"answers": ["src/chunkflow/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 mo...
ctx_compress-s1-00808-twin
chunkflow
ctx_compress-s1-00808
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_lexeme raises ValueError when a value itself contains an equals sign. It lives in src/chunkflow/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 chunkflow l...
22,001
1,100
44
87
false
synth:74515841235785874@1
synth
[{"aliases": ["src/chunkflow/format.py", "format.py"], "kind": "file", "value": "src/chunkflow/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["src/chunkflow/compat.py", "compat.py"], "kind": "file", "value": "src/chunkflow/compat.py"},...
[{"answers": ["src/chunkflow/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 mo...
ctx_compress-s1-00809
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] net_invoice raises ReceiptRangeError for bytes input in strict mode instead of decoding it. It lives in src/invoicedesk/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 invo...
51,706
2,585
78
97
false
synth:6132374930909523698@1
synth
[{"aliases": ["src/invoicedesk/io.py", "io.py"], "kind": "file", "value": "src/invoicedesk/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["docs/changelog.md", "changelog.md"], "kind": "file", "value": "docs/CHANGELOG.md"}, {"aliases": ["data/balance_4...
[{"answers": ["src/invoicedesk/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-s1-00809-twin
invoicedesk
ctx_compress-s1-00809
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] net_invoice raises ReceiptRangeError for bytes input in strict mode instead of decoding it. It lives in src/invoicedesk/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 invo...
42,341
2,117
54
97
false
synth:6132374930909523698@1
synth
[{"aliases": ["src/invoicedesk/io.py", "io.py"], "kind": "file", "value": "src/invoicedesk/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["docs/changelog.md", "changelog.md"], "kind": "file", "value": "docs/CHANGELOG.md"}, {"aliases": ["data/balance_4...
[{"answers": ["src/invoicedesk/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-s1-00810
palletkit
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 keeps leading spaces in keys so lookups after a semicolon fail. It lives in src/palletkit/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 palletkit layout....
53,329
2,666
73
78
false
synth:5461274478698806812@1
synth
[{"aliases": ["src/palletkit/cli.py", "cli.py"], "kind": "file", "value": "src/palletkit/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["data/pallet_1.csv", "pallet_1.csv"], "kind": "file", "value": "data/pallet_1.csv"}, {"aliases": ["src/palletki...
[{"answers": ["src/palletkit/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 rec...
ctx_compress-s1-00810-twin
palletkit
ctx_compress-s1-00810
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 keeps leading spaces in keys so lookups after a semicolon fail. It lives in src/palletkit/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 palletkit layout....
31,141
1,557
43
78
false
synth:5461274478698806812@1
synth
[{"aliases": ["src/palletkit/cli.py", "cli.py"], "kind": "file", "value": "src/palletkit/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["data/pallet_1.csv", "pallet_1.csv"], "kind": "file", "value": "data/pallet_1.csv"}, {"aliases": ["src/palletki...
[{"answers": ["src/palletkit/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 rec...
ctx_compress-s1-00811
waypointforge
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] simplify_zone_all raises CorridorDriftError for bytes input in strict mode instead of decoding it. It lives in src/waypointforge/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...
23,207
1,160
39
44
false
synth:6138309853350652168@1
synth
[{"aliases": ["src/waypointforge/report.py", "report.py"], "kind": "file", "value": "src/waypointforge/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["src/waypointforge/tables.py", "tables.py"], "kind": "file", "value": "src/waypointfo...
[{"answers": ["src/waypointforge/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'...
ctx_compress-s1-00811-twin
waypointforge
ctx_compress-s1-00811
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] simplify_zone_all raises CorridorDriftError for bytes input in strict mode instead of decoding it. It lives in src/waypointforge/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...
18,324
1,000
25
44
false
synth:6138309853350652168@1
synth
[{"aliases": ["src/waypointforge/report.py", "report.py"], "kind": "file", "value": "src/waypointforge/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["src/waypointforge/tables.py", "tables.py"], "kind": "file", "value": "src/waypointfo...
[{"answers": ["no"], "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 test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["test_cli.py", "tests/test_cli.py"], "...
ctx_compress-s1-00812
captionledger
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_caption raises KeyError for a missing key instead of returning the default. It lives in src/captionledger/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 ...
32,923
1,646
62
79
false
synth:5572364813089901495@1
synth
[{"aliases": ["src/captionledger/report.py", "report.py"], "kind": "file", "value": "src/captionledger/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["data/slug_1.csv", "slug_1.csv"], "kind": "file", "value": "data/slug_1.csv"}, {"alia...
[{"answers": ["src/captionledger/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'...
ctx_compress-s1-00812-twin
captionledger
ctx_compress-s1-00812
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_caption raises KeyError for a missing key instead of returning the default. It lives in src/captionledger/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 ...
14,209
1,000
37
79
false
synth:5572364813089901495@1
synth
[{"aliases": ["src/captionledger/report.py", "report.py"], "kind": "file", "value": "src/captionledger/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["data/slug_1.csv", "slug_1.csv"], "kind": "file", "value": "data/slug_1.csv"}, {"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": ["audit.py", "src/captionledger/audit.p...
ctx_compress-s1-00813
legforge
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] cluster_waypoint double counts carried amounts when the period closes on a weekend. It lives in src/legforge/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 legforg...
56,558
2,827
75
80
false
synth:838516757957565197@1
synth
[{"aliases": ["src/legforge/format.py", "format.py"], "kind": "file", "value": "src/legforge/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["src/legforge/schema.py", "schema.py"], "kind": "file", "value": "src/legforge/schema.py"}, {"a...
[{"answers": ["src/legforge/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 mos...
ctx_compress-s1-00813-twin
legforge
ctx_compress-s1-00813
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] cluster_waypoint double counts carried amounts when the period closes on a weekend. It lives in src/legforge/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 legforg...
25,137
1,256
41
80
false
synth:838516757957565197@1
synth
[{"aliases": ["src/legforge/format.py", "format.py"], "kind": "file", "value": "src/legforge/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["src/legforge/schema.py", "schema.py"], "kind": "file", "value": "src/legforge/schema.py"}, {"a...
[{"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/leg_2.csv", "leg_2.csv"], "kind"...
ctx_compress-s1-00814
spankit
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] bin_span double counts carried amounts when the period closes on a weekend. It lives in src/spankit/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 spankit layout. [t...
42,772
2,138
68
83
false
synth:3843217392271401698@1
synth
[{"aliases": ["src/spankit/audit.py", "audit.py"], "kind": "file", "value": "src/spankit/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["src/spankit/cli.py", "cli.py"], "kind": "file", "value": "src/spankit/cli.py"}, {"aliases": ["data/gau...
[{"answers": ["src/spankit/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 most r...
ctx_compress-s1-00814-twin
spankit
ctx_compress-s1-00814
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] bin_span double counts carried amounts when the period closes on a weekend. It lives in src/spankit/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 spankit layout. [t...
28,237
1,411
40
83
false
synth:3843217392271401698@1
synth
[{"aliases": ["src/spankit/audit.py", "audit.py"], "kind": "file", "value": "src/spankit/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["src/spankit/cli.py", "cli.py"], "kind": "file", "value": "src/spankit/cli.py"}, {"aliases": ["data/gau...
[{"answers": ["src/spankit/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 most r...
ctx_compress-s1-00815
sectionledger
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] anchor_clause raises ValueError when a value itself contains an equals sign. It lives in src/sectionledger/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 sectionledger lay...
68,582
3,072
81
100
false
synth:20420878345896638@2.11675
synth
[{"aliases": ["src/sectionledger/io.py", "io.py"], "kind": "file", "value": "src/sectionledger/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["src/sectionledger/tables.py", "tables.py"], "kind": "file", "value": "src/sectionledger/tables.py"}, {"alias...
[{"answers": ["src/sectionledger/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 r...
ctx_compress-s1-00815-twin
sectionledger
ctx_compress-s1-00815
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] anchor_clause raises ValueError when a value itself contains an equals sign. It lives in src/sectionledger/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 sectionledger lay...
32,647
1,632
47
100
false
synth:20420878345896638@2.11675
synth
[{"aliases": ["src/sectionledger/io.py", "io.py"], "kind": "file", "value": "src/sectionledger/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["src/sectionledger/tables.py", "tables.py"], "kind": "file", "value": "src/sectionledger/tables.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": ["src/sectionledger/tables.py", "tables...
ctx_compress-s1-00816
scopeledger
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 raises ValueError when a value itself contains an equals sign. It lives in src/scopeledger/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 scopeledger la...
100,468
3,072
103
116
false
synth:6529196117166132802@1.76487
synth
[{"aliases": ["src/scopeledger/compat.py", "compat.py"], "kind": "file", "value": "src/scopeledger/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["data/policy_4.csv", "policy_4.csv"], "kind": "file", "value": "data/policy_4.csv"}, {"al...
[{"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-s1-00816-twin
scopeledger
ctx_compress-s1-00816
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 raises ValueError when a value itself contains an equals sign. It lives in src/scopeledger/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 scopeledger la...
52,686
2,634
59
116
false
synth:6529196117166132802@1.76487
synth
[{"aliases": ["src/scopeledger/compat.py", "compat.py"], "kind": "file", "value": "src/scopeledger/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["data/policy_4.csv", "policy_4.csv"], "kind": "file", "value": "data/policy_4.csv"}, {"al...
[{"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/session_9.csv", "session_9.csv"]...
ctx_compress-s1-00817
balancekit
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] settle_ledger_safe returns the upper bound for every value instead of clamping into the range. It lives in src/balancekit/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 ...
20,993
1,049
41
48
false
synth:5483241058551634501@1
synth
[{"aliases": ["src/balancekit/cli.py", "cli.py"], "kind": "file", "value": "src/balancekit/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["data/tariff_6.csv", "tariff_6.csv"], "kind": "file", "value": "data/tariff_6.csv"}, {"aliases": ["src/balanc...
[{"answers": ["src/balancekit/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-s1-00817-twin
balancekit
ctx_compress-s1-00817
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] settle_ledger_safe returns the upper bound for every value instead of clamping into the range. It lives in src/balancekit/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 ...
13,309
1,000
27
48
false
synth:5483241058551634501@1
synth
[{"aliases": ["src/balancekit/cli.py", "cli.py"], "kind": "file", "value": "src/balancekit/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["data/tariff_6.csv", "tariff_6.csv"], "kind": "file", "value": "data/tariff_6.csv"}, {"aliases": ["src/balanc...
[{"answers": ["no"], "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 test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["data/refund_4.csv", "refund_4.csv"], ...
ctx_compress-s1-00818
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_fix keeps leading spaces in keys so lookups after a semicolon fail. It lives in src/beaconledger/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 beaconledger la...
19,216
1,000
49
54
false
synth:1118275220353140474@1
synth
[{"aliases": ["src/beaconledger/io.py", "io.py"], "kind": "file", "value": "src/beaconledger/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["src/beaconledger/store.py", "store.py"], "kind": "file", "value": "src/beaconledger/store.py"}, {"aliases": ["...
[{"answers": ["src/beaconledger/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 re...
ctx_compress-s1-00818-twin
beaconledger
ctx_compress-s1-00818
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_fix keeps leading spaces in keys so lookups after a semicolon fail. It lives in src/beaconledger/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 beaconledger la...
10,302
1,000
25
54
false
synth:1118275220353140474@1
synth
[{"aliases": ["src/beaconledger/io.py", "io.py"], "kind": "file", "value": "src/beaconledger/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["src/beaconledger/store.py", "store.py"], "kind": "file", "value": "src/beaconledger/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": ["test_format.py", "tests/test_format.p...
ctx_compress-s1-00819
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_cue keeps leading spaces in keys so lookups after a semicolon fail. It lives in src/slugdesk/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 slugdesk layout. [t...
103,343
3,072
100
119
false
synth:6099501759237325329@1.7362
synth
[{"aliases": ["src/slugdesk/cli.py", "cli.py"], "kind": "file", "value": "src/slugdesk/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["data/stem_4.csv", "stem_4.csv"], "kind": "file", "value": "data/stem_4.csv"}, {"aliases": ["data/chunk_10.csv", ...
[{"answers": ["src/slugdesk/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-s1-00819-twin
slugdesk
ctx_compress-s1-00819
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_cue keeps leading spaces in keys so lookups after a semicolon fail. It lives in src/slugdesk/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 slugdesk layout. [t...
57,538
2,876
66
119
false
synth:6099501759237325329@1.7362
synth
[{"aliases": ["src/slugdesk/cli.py", "cli.py"], "kind": "file", "value": "src/slugdesk/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["data/stem_4.csv", "stem_4.csv"], "kind": "file", "value": "data/stem_4.csv"}, {"aliases": ["data/chunk_10.csv", ...
[{"answers": ["src/slugdesk/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-s1-00820
reorderdesk
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_lot_safe drops override values of 0, False and empty string. It lives in src/reorderdesk/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 reorderdesk layout. [...
67,119
3,072
103
110
false
synth:642561808113646039@1.85308
synth
[{"aliases": ["src/reorderdesk/rules.py", "rules.py"], "kind": "file", "value": "src/reorderdesk/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/reorderdesk/cli.py", "cli.py"], "kind": "file", "value": "src/reorderdesk/cli.py"}, {"alia...
[{"answers": ["src/reorderdesk/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-s1-00820-twin
reorderdesk
ctx_compress-s1-00820
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_lot_safe drops override values of 0, False and empty string. It lives in src/reorderdesk/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 reorderdesk layout. [...
38,747
1,937
55
110
false
synth:642561808113646039@1.85308
synth
[{"aliases": ["src/reorderdesk/rules.py", "rules.py"], "kind": "file", "value": "src/reorderdesk/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/reorderdesk/cli.py", "cli.py"], "kind": "file", "value": "src/reorderdesk/cli.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": ["test_cli.py", "tests/test_cli.py"], "...
ctx_compress-s1-00821
cueworks
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/cueworks/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 cueworks layout. [tool_use kind=searc...
50,639
2,531
73
78
false
synth:1356808786870691026@1
synth
[{"aliases": ["src/cueworks/limits.py", "limits.py"], "kind": "file", "value": "src/cueworks/limits.py"}, {"aliases": ["tests/test_limits.py", "test_limits.py"], "kind": "file", "value": "tests/test_limits.py"}, {"aliases": ["src/cueworks/tables.py", "tables.py"], "kind": "file", "value": "src/cueworks/tables.py"}, {"a...
[{"answers": ["src/cueworks/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 mos...
ctx_compress-s1-00821-twin
cueworks
ctx_compress-s1-00821
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/cueworks/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 cueworks layout. [tool_use kind=searc...
26,988
1,349
45
78
false
synth:1356808786870691026@1
synth
[{"aliases": ["src/cueworks/limits.py", "limits.py"], "kind": "file", "value": "src/cueworks/limits.py"}, {"aliases": ["tests/test_limits.py", "test_limits.py"], "kind": "file", "value": "tests/test_limits.py"}, {"aliases": ["src/cueworks/tables.py", "tables.py"], "kind": "file", "value": "src/cueworks/tables.py"}, {"a...
[{"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/cueworks/store.py", "store.py"], ...
ctx_compress-s1-00822
sampleflow
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_bucket_once keeps leading spaces in keys so lookups after a semicolon fail. It lives in src/sampleflow/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 samplefl...
38,457
1,922
101
118
false
synth:2861061226747236473@1.24415
synth
[{"aliases": ["src/sampleflow/limits.py", "limits.py"], "kind": "file", "value": "src/sampleflow/limits.py"}, {"aliases": ["tests/test_limits.py", "test_limits.py"], "kind": "file", "value": "tests/test_limits.py"}, {"aliases": ["src/sampleflow/rules.py", "rules.py"], "kind": "file", "value": "src/sampleflow/rules.py"}...
[{"answers": ["src/sampleflow/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 m...
ctx_compress-s1-00822-twin
sampleflow
ctx_compress-s1-00822
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_bucket_once keeps leading spaces in keys so lookups after a semicolon fail. It lives in src/sampleflow/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 samplefl...
19,281
1,000
55
118
false
synth:2861061226747236473@1.24415
synth
[{"aliases": ["src/sampleflow/limits.py", "limits.py"], "kind": "file", "value": "src/sampleflow/limits.py"}, {"aliases": ["tests/test_limits.py", "test_limits.py"], "kind": "file", "value": "tests/test_limits.py"}, {"aliases": ["src/sampleflow/rules.py", "rules.py"], "kind": "file", "value": "src/sampleflow/rules.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": ["src/sampleflow/tables.py", "tables.py...
ctx_compress-s1-00823
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] split_quota rejects a range where lo equals hi, which is a valid single point range. It lives in src/rosterledger/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 roster...
50,953
2,547
92
115
false
synth:819787153268724900@1.14015
synth
[{"aliases": ["src/rosterledger/core.py", "core.py"], "kind": "file", "value": "src/rosterledger/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/rosterledger/cli.py", "cli.py"], "kind": "file", "value": "src/rosterledger/cli.py"}, {"aliase...
[{"answers": ["src/rosterledger/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": ["pass"], "question": "Did the agent's mos...
ctx_compress-s1-00823-twin
rosterledger
ctx_compress-s1-00823
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] split_quota rejects a range where lo equals hi, which is a valid single point range. It lives in src/rosterledger/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 roster...
30,384
1,519
60
115
false
synth:819787153268724900@1.14015
synth
[{"aliases": ["src/rosterledger/core.py", "core.py"], "kind": "file", "value": "src/rosterledger/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/rosterledger/cli.py", "cli.py"], "kind": "file", "value": "src/rosterledger/cli.py"}, {"aliase...
[{"answers": ["src/rosterledger/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 mos...
ctx_compress-s1-00824
indexdesk
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_heading keeps leading spaces in keys so lookups after a semicolon fail. It lives in src/indexdesk/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 indexdesk lay...
118,994
3,072
133
146
false
synth:288402712998837893@2.1464
synth
[{"aliases": ["src/indexdesk/core.py", "core.py"], "kind": "file", "value": "src/indexdesk/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["data/index_13.csv", "index_13.csv"], "kind": "file", "value": "data/index_13.csv"}, {"aliases": ["data/h...
[{"answers": ["src/indexdesk/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": ["pass"], "question": "Did the agent's most r...
ctx_compress-s1-00824-twin
indexdesk
ctx_compress-s1-00824
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_heading keeps leading spaces in keys so lookups after a semicolon fail. It lives in src/indexdesk/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 indexdesk lay...
69,274
3,072
79
146
false
synth:288402712998837893@2.1464
synth
[{"aliases": ["src/indexdesk/core.py", "core.py"], "kind": "file", "value": "src/indexdesk/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["data/index_13.csv", "index_13.csv"], "kind": "file", "value": "data/index_13.csv"}, {"aliases": ["data/h...
[{"answers": ["src/indexdesk/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": ["pass"], "question": "Did the agent's most r...
ctx_compress-s1-00825
sectionflow
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] collate_section_page keeps leading spaces in keys so lookups after a semicolon fail. It lives in src/sectionflow/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 secti...
56,828
2,841
95
102
false
synth:5936999109135975300@1
synth
[{"aliases": ["src/sectionflow/audit.py", "audit.py"], "kind": "file", "value": "src/sectionflow/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["data/clause_1.csv", "clause_1.csv"], "kind": "file", "value": "data/clause_1.csv"}, {"aliases"...
[{"answers": ["src/sectionflow/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 mo...
ctx_compress-s1-00825-twin
sectionflow
ctx_compress-s1-00825
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] collate_section_page keeps leading spaces in keys so lookups after a semicolon fail. It lives in src/sectionflow/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 secti...
25,230
1,261
53
102
false
synth:5936999109135975300@1
synth
[{"aliases": ["src/sectionflow/audit.py", "audit.py"], "kind": "file", "value": "src/sectionflow/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["data/clause_1.csv", "clause_1.csv"], "kind": "file", "value": "data/clause_1.csv"}, {"aliases"...
[{"answers": ["src/sectionflow/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 mo...
ctx_compress-s1-00826
tariffworks
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_receipt lets defaults overwrite explicit overrides. It lives in src/tariffworks/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 tariffworks layout. [tool_use...
67,413
3,072
92
107
false
synth:8988269670683599864@1.13815
synth
[{"aliases": ["src/tariffworks/report.py", "report.py"], "kind": "file", "value": "src/tariffworks/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["data/balance_4.csv", "balance_4.csv"], "kind": "file", "value": "data/balance_4.csv"}, {...
[{"answers": ["src/tariffworks/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-s1-00826-twin
tariffworks
ctx_compress-s1-00826
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_receipt lets defaults overwrite explicit overrides. It lives in src/tariffworks/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 tariffworks layout. [tool_use...
34,961
1,748
50
107
false
synth:8988269670683599864@1.13815
synth
[{"aliases": ["src/tariffworks/report.py", "report.py"], "kind": "file", "value": "src/tariffworks/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["data/balance_4.csv", "balance_4.csv"], "kind": "file", "value": "data/balance_4.csv"}, {...
[{"answers": ["src/tariffworks/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-s1-00827
revisionledger
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] renumber_revision drops override values of 0, False and empty string. It lives in src/revisionledger/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 revisionledger ...
74,309
3,072
82
99
false
synth:8276959761860733990@1.51798
synth
[{"aliases": ["src/revisionledger/compat.py", "compat.py"], "kind": "file", "value": "src/revisionledger/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["data/clause_8.csv", "clause_8.csv"], "kind": "file", "value": "data/clause_8.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": ["data/footnote_3.csv", "footnote_3.csv...
ctx_compress-s1-00827-twin
revisionledger
ctx_compress-s1-00827
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] renumber_revision drops override values of 0, False and empty string. It lives in src/revisionledger/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 revisionledger ...
48,595
2,429
56
99
false
synth:8276959761860733990@1.51798
synth
[{"aliases": ["src/revisionledger/compat.py", "compat.py"], "kind": "file", "value": "src/revisionledger/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["data/clause_8.csv", "clause_8.csv"], "kind": "file", "value": "data/clause_8.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": ["data/index_5.csv", "index_5.csv"], "k...
ctx_compress-s1-00828
lotforge
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_sku returns one item too few on every page. It lives in src/lotforge/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 lotforge layout. [tool_use kind=search nam...
113,702
3,072
151
166
false
synth:8228546417670805254@2.19592
synth
[{"aliases": ["src/lotforge/report.py", "report.py"], "kind": "file", "value": "src/lotforge/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["data/stock_3.csv", "stock_3.csv"], "kind": "file", "value": "data/stock_3.csv"}, {"aliases": [...
[{"answers": ["src/lotforge/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 mos...
ctx_compress-s1-00828-twin
lotforge
ctx_compress-s1-00828
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_sku returns one item too few on every page. It lives in src/lotforge/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 lotforge layout. [tool_use kind=search nam...
80,742
3,072
91
166
false
synth:8228546417670805254@2.19592
synth
[{"aliases": ["src/lotforge/report.py", "report.py"], "kind": "file", "value": "src/lotforge/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["data/stock_3.csv", "stock_3.csv"], "kind": "file", "value": "data/stock_3.csv"}, {"aliases": [...
[{"answers": ["src/lotforge/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 mos...
ctx_compress-s1-00829
tariffkit
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 raises ValueError when a value itself contains an equals sign. It lives in src/tariffkit/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 tariffkit layout....
53,020
2,651
106
115
false
synth:9151548427718379111@1.44097
synth
[{"aliases": ["src/tariffkit/io.py", "io.py"], "kind": "file", "value": "src/tariffkit/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["src/tariffkit/tables.py", "tables.py"], "kind": "file", "value": "src/tariffkit/tables.py"}, {"aliases": ["src/tarif...
[{"answers": ["src/tariffkit/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 recen...
ctx_compress-s1-00829-twin
tariffkit
ctx_compress-s1-00829
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 raises ValueError when a value itself contains an equals sign. It lives in src/tariffkit/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 tariffkit layout....
36,446
1,822
60
115
false
synth:9151548427718379111@1.44097
synth
[{"aliases": ["src/tariffkit/io.py", "io.py"], "kind": "file", "value": "src/tariffkit/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["src/tariffkit/tables.py", "tables.py"], "kind": "file", "value": "src/tariffkit/tables.py"}, {"aliases": ["src/tarif...
[{"answers": ["src/tariffkit/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 recen...
ctx_compress-s1-00830
skudesk
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_lot misses keys that differ only in case. It lives in src/skudesk/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 skudesk layout. [tool_use kind=search name=Bash...
64,720
3,072
106
119
false
synth:150474428363924858@2.00215
synth
[{"aliases": ["src/skudesk/store.py", "store.py"], "kind": "file", "value": "src/skudesk/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/skudesk/core.py", "core.py"], "kind": "file", "value": "src/skudesk/core.py"}, {"aliases": ["data/...
[{"answers": ["src/skudesk/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 most r...
ctx_compress-s1-00830-twin
skudesk
ctx_compress-s1-00830
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_lot misses keys that differ only in case. It lives in src/skudesk/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 skudesk layout. [tool_use kind=search name=Bash...
38,513
1,925
63
119
false
synth:150474428363924858@2.00215
synth
[{"aliases": ["src/skudesk/store.py", "store.py"], "kind": "file", "value": "src/skudesk/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/skudesk/core.py", "core.py"], "kind": "file", "value": "src/skudesk/core.py"}, {"aliases": ["data/...
[{"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/skudesk/schema.py"],...
ctx_compress-s1-00831
bearingkit
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] bound_waypoint raises ZeroDivisionError on the first value and lags by one afterwards. It lives in src/bearingkit/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 bear...
95,610
3,072
101
114
false
synth:9155289457478995726@2.15923
synth
[{"aliases": ["src/bearingkit/store.py", "store.py"], "kind": "file", "value": "src/bearingkit/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["data/beacon_4.csv", "beacon_4.csv"], "kind": "file", "value": "data/beacon_4.csv"}, {"aliases": ...
[{"answers": ["src/bearingkit/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-s1-00831-twin
bearingkit
ctx_compress-s1-00831
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] bound_waypoint raises ZeroDivisionError on the first value and lags by one afterwards. It lives in src/bearingkit/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 bear...
51,201
2,560
65
114
false
synth:9155289457478995726@2.15923
synth
[{"aliases": ["src/bearingkit/store.py", "store.py"], "kind": "file", "value": "src/bearingkit/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["data/beacon_4.csv", "beacon_4.csv"], "kind": "file", "value": "data/beacon_4.csv"}, {"aliases": ...
[{"answers": ["src/bearingkit/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-s1-00832
slugflow
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_snippet_all raises ZeroDivisionError on the first value and lags by one afterwards. It lives in src/slugflow/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 slugflow...
29,329
1,466
59
68
false
synth:2487489048748840820@1
synth
[{"aliases": ["src/slugflow/io.py", "io.py"], "kind": "file", "value": "src/slugflow/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["src/slugflow/store.py", "store.py"], "kind": "file", "value": "src/slugflow/store.py"}, {"aliases": ["src/slugflow/aud...
[{"answers": ["src/slugflow/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 recent...
ctx_compress-s1-00832-twin
slugflow
ctx_compress-s1-00832
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_snippet_all raises ZeroDivisionError on the first value and lags by one afterwards. It lives in src/slugflow/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 slugflow...
17,424
1,000
37
68
false
synth:2487489048748840820@1
synth
[{"aliases": ["src/slugflow/io.py", "io.py"], "kind": "file", "value": "src/slugflow/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["src/slugflow/store.py", "store.py"], "kind": "file", "value": "src/slugflow/store.py"}, {"aliases": ["src/slugflow/aud...
[{"answers": ["src/slugflow/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-s1-00833
sessionkit
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_nonce returns the upper bound for every value instead of clamping into the range. It lives in src/sessionkit/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 sessio...
78,120
3,072
99
124
false
synth:4319246899874804481@1.77757
synth
[{"aliases": ["src/sessionkit/core.py", "core.py"], "kind": "file", "value": "src/sessionkit/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["data/grant_13.csv", "grant_13.csv"], "kind": "file", "value": "data/grant_13.csv"}, {"aliases": ["data...
[{"answers": ["src/sessionkit/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": ["pass"], "question": "Did the agent's most ...
ctx_compress-s1-00833-twin
sessionkit
ctx_compress-s1-00833
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_nonce returns the upper bound for every value instead of clamping into the range. It lives in src/sessionkit/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 sessio...
50,781
2,539
61
124
false
synth:4319246899874804481@1.77757
synth
[{"aliases": ["src/sessionkit/core.py", "core.py"], "kind": "file", "value": "src/sessionkit/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["data/grant_13.csv", "grant_13.csv"], "kind": "file", "value": "data/grant_13.csv"}, {"aliases": ["data...
[{"answers": ["src/sessionkit/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-s1-00834
counterflow
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] smooth_trace raises TraceShapeError for bytes input in strict mode instead of decoding it. It lives in src/counterflow/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 count...
35,928
1,796
62
75
false
synth:387528054804012958@1
synth
[{"aliases": ["src/counterflow/io.py", "io.py"], "kind": "file", "value": "src/counterflow/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["src/counterflow/schema.py", "schema.py"], "kind": "file", "value": "src/counterflow/schema.py"}, {"aliases": ["d...
[{"answers": ["src/counterflow/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-s1-00834-twin
counterflow
ctx_compress-s1-00834
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] smooth_trace raises TraceShapeError for bytes input in strict mode instead of decoding it. It lives in src/counterflow/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 count...
19,433
1,000
42
75
false
synth:387528054804012958@1
synth
[{"aliases": ["src/counterflow/io.py", "io.py"], "kind": "file", "value": "src/counterflow/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["src/counterflow/schema.py", "schema.py"], "kind": "file", "value": "src/counterflow/schema.py"}, {"aliases": ["d...
[{"answers": ["src/counterflow/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-s1-00835
nonceflow
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_grant_page returns the upper bound for every value instead of clamping into the range. It lives in src/nonceflow/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 th...
40,937
2,046
62
79
false
synth:1744588170602705548@1
synth
[{"aliases": ["src/nonceflow/rules.py", "rules.py"], "kind": "file", "value": "src/nonceflow/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/nonceflow/store.py", "store.py"], "kind": "file", "value": "src/nonceflow/store.py"}, {"aliase...
[{"answers": ["src/nonceflow/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...
ctx_compress-s1-00835-twin
nonceflow
ctx_compress-s1-00835
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_grant_page returns the upper bound for every value instead of clamping into the range. It lives in src/nonceflow/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 th...
22,028
1,101
42
79
false
synth:1744588170602705548@1
synth
[{"aliases": ["src/nonceflow/rules.py", "rules.py"], "kind": "file", "value": "src/nonceflow/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/nonceflow/store.py", "store.py"], "kind": "file", "value": "src/nonceflow/store.py"}, {"aliase...
[{"answers": ["src/nonceflow/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 most...
ctx_compress-s1-00836
revisionledger
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_footnote averages one value too many once the window is full. It lives in src/revisionledger/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 revisionledger layout....
68,242
3,072
100
127
false
synth:647740309814242376@1.6951
synth
[{"aliases": ["src/revisionledger/core.py", "core.py"], "kind": "file", "value": "src/revisionledger/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/revisionledger/format.py", "format.py"], "kind": "file", "value": "src/revisionledger/form...
[{"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-s1-00836-twin
revisionledger
ctx_compress-s1-00836
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_footnote averages one value too many once the window is full. It lives in src/revisionledger/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 revisionledger layout....
44,966
2,248
72
127
false
synth:647740309814242376@1.6951
synth
[{"aliases": ["src/revisionledger/core.py", "core.py"], "kind": "file", "value": "src/revisionledger/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/revisionledger/format.py", "format.py"], "kind": "file", "value": "src/revisionledger/form...
[{"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/footnote_4.csv", "footnote_4.csv...
ctx_compress-s1-00837
shiftflow
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_calendar_page double counts carried amounts when the period closes on a weekend. It lives in src/shiftflow/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 shi...
39,409
1,970
64
77
false
synth:4487113026566776096@1
synth
[{"aliases": ["src/shiftflow/rules.py", "rules.py"], "kind": "file", "value": "src/shiftflow/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/shiftflow/tables.py", "tables.py"], "kind": "file", "value": "src/shiftflow/tables.py"}, {"ali...
[{"answers": ["src/shiftflow/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...
ctx_compress-s1-00837-twin
shiftflow
ctx_compress-s1-00837
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_calendar_page double counts carried amounts when the period closes on a weekend. It lives in src/shiftflow/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 shi...
22,437
1,121
42
77
false
synth:4487113026566776096@1
synth
[{"aliases": ["src/shiftflow/rules.py", "rules.py"], "kind": "file", "value": "src/shiftflow/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/shiftflow/tables.py", "tables.py"], "kind": "file", "value": "src/shiftflow/tables.py"}, {"ali...
[{"answers": ["src/shiftflow/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 most...
ctx_compress-s1-00838
clausekit
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] renumber_index leaves a trailing dash when the title ends in punctuation. It lives in src/clausekit/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 clausekit layout. ...
28,890
1,444
68
73
false
synth:5763488378987264636@1
synth
[{"aliases": ["src/clausekit/rules.py", "rules.py"], "kind": "file", "value": "src/clausekit/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["docs/changelog.md", "changelog.md"], "kind": "file", "value": "docs/CHANGELOG.md"}, {"aliases": ["...
[{"answers": ["src/clausekit/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 most...
ctx_compress-s1-00838-twin
clausekit
ctx_compress-s1-00838
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] renumber_index leaves a trailing dash when the title ends in punctuation. It lives in src/clausekit/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 clausekit layout. ...
19,181
1,000
42
73
false
synth:5763488378987264636@1
synth
[{"aliases": ["src/clausekit/rules.py", "rules.py"], "kind": "file", "value": "src/clausekit/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["docs/changelog.md", "changelog.md"], "kind": "file", "value": "docs/CHANGELOG.md"}, {"aliases": ["...
[{"answers": ["src/clausekit/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...
ctx_compress-s1-00839
balanceworks
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_once returns one item too few on every page. It lives in src/balanceworks/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 balanceworks layout. [tool...
21,807
1,090
52
59
false
synth:8862213370284448913@1
synth
[{"aliases": ["src/balanceworks/store.py", "store.py"], "kind": "file", "value": "src/balanceworks/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/balanceworks/io.py", "io.py"], "kind": "file", "value": "src/balanceworks/io.py"}, {"ali...
[{"answers": ["src/balanceworks/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 m...
ctx_compress-s1-00839-twin
balanceworks
ctx_compress-s1-00839
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_once returns one item too few on every page. It lives in src/balanceworks/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 balanceworks layout. [tool...
11,511
1,000
30
59
false
synth:8862213370284448913@1
synth
[{"aliases": ["src/balanceworks/store.py", "store.py"], "kind": "file", "value": "src/balanceworks/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/balanceworks/io.py", "io.py"], "kind": "file", "value": "src/balanceworks/io.py"}, {"ali...
[{"answers": ["src/balanceworks/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 m...
ctx_compress-s1-00840
tiledesk
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] cluster_bearing rejects a range where lo equals hi, which is a valid single point range. It lives in src/tiledesk/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 tiledesk...
33,438
1,671
60
75
false
synth:1635544344459234258@1
synth
[{"aliases": ["src/tiledesk/cli.py", "cli.py"], "kind": "file", "value": "src/tiledesk/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["data/beacon_1.csv", "beacon_1.csv"], "kind": "file", "value": "data/beacon_1.csv"}, {"aliases": ["src/tiledesk/t...
[{"answers": ["src/tiledesk/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-s1-00840-twin
tiledesk
ctx_compress-s1-00840
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] cluster_bearing rejects a range where lo equals hi, which is a valid single point range. It lives in src/tiledesk/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 tiledesk...
16,249
1,000
34
75
false
synth:1635544344459234258@1
synth
[{"aliases": ["src/tiledesk/cli.py", "cli.py"], "kind": "file", "value": "src/tiledesk/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["data/beacon_1.csv", "beacon_1.csv"], "kind": "file", "value": "data/beacon_1.csv"}, {"aliases": ["src/tiledesk/t...
[{"answers": ["src/tiledesk/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-s1-00841
balanceforge
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] net_tariff returns the upper bound for every value instead of clamping into the range. It lives in src/balanceforge/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 ba...
48,330
2,416
98
125
false
synth:1323556197939870712@2.0746
synth
[{"aliases": ["src/balanceforge/rules.py", "rules.py"], "kind": "file", "value": "src/balanceforge/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/balanceforge/audit.py", "audit.py"], "kind": "file", "value": "src/balanceforge/audit.py...
[{"answers": ["src/balanceforge/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 m...
ctx_compress-s1-00841-twin
balanceforge
ctx_compress-s1-00841
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] net_tariff returns the upper bound for every value instead of clamping into the range. It lives in src/balanceforge/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 ba...
42,155
2,107
69
125
false
synth:1323556197939870712@2.0746
synth
[{"aliases": ["src/balanceforge/rules.py", "rules.py"], "kind": "file", "value": "src/balanceforge/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/balanceforge/audit.py", "audit.py"], "kind": "file", "value": "src/balanceforge/audit.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": ["compat.py", "src/balanceforge/compat....
ctx_compress-s1-00842
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] roll_ledger_all produces doubled dashes for consecutive separators. It lives in src/ledgerdesk/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 ledgerdesk layout. [tool_use ...
109,080
3,072
118
131
false
synth:2510584474493613546@1.86553
synth
[{"aliases": ["src/ledgerdesk/io.py", "io.py"], "kind": "file", "value": "src/ledgerdesk/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["src/ledgerdesk/cli.py", "cli.py"], "kind": "file", "value": "src/ledgerdesk/cli.py"}, {"aliases": ["data/ledger_2....
[{"answers": ["src/ledgerdesk/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 rece...
ctx_compress-s1-00842-twin
ledgerdesk
ctx_compress-s1-00842
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_ledger_all produces doubled dashes for consecutive separators. It lives in src/ledgerdesk/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 ledgerdesk layout. [tool_use ...
54,057
2,702
66
131
false
synth:2510584474493613546@1.86553
synth
[{"aliases": ["src/ledgerdesk/io.py", "io.py"], "kind": "file", "value": "src/ledgerdesk/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["src/ledgerdesk/cli.py", "cli.py"], "kind": "file", "value": "src/ledgerdesk/cli.py"}, {"aliases": ["data/ledger_2....
[{"answers": ["src/ledgerdesk/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 rece...
ctx_compress-s1-00843
cueworks
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_cue ignores the cap, so late attempts wait for minutes. It lives in src/cueworks/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 cueworks layout. [tool_us...
51,714
2,585
86
103
false
synth:1191308574097013160@1.18893
synth
[{"aliases": ["src/cueworks/compat.py", "compat.py"], "kind": "file", "value": "src/cueworks/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["docs/changelog.md", "changelog.md"], "kind": "file", "value": "docs/CHANGELOG.md"}, {"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_format.py", "tests/test_format.p...
ctx_compress-s1-00843-twin
cueworks
ctx_compress-s1-00843
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_cue ignores the cap, so late attempts wait for minutes. It lives in src/cueworks/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 cueworks layout. [tool_us...
27,160
1,358
56
103
false
synth:1191308574097013160@1.18893
synth
[{"aliases": ["src/cueworks/compat.py", "compat.py"], "kind": "file", "value": "src/cueworks/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["docs/changelog.md", "changelog.md"], "kind": "file", "value": "docs/CHANGELOG.md"}, {"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": ["changelog.md", "docs/changelog.md"], ...
ctx_compress-s1-00844
binflow
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_warehouse misses keys that differ only in case. It lives in src/binflow/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 binflow layout. [tool_use kind=search n...
25,053
1,252
55
60
false
synth:6485255560750804075@1
synth
[{"aliases": ["src/binflow/core.py", "core.py"], "kind": "file", "value": "src/binflow/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/binflow/audit.py", "audit.py"], "kind": "file", "value": "src/binflow/audit.py"}, {"aliases": ["src/binf...
[{"answers": ["src/binflow/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 rec...
ctx_compress-s1-00844-twin
binflow
ctx_compress-s1-00844
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_warehouse misses keys that differ only in case. It lives in src/binflow/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 binflow layout. [tool_use kind=search n...
9,975
1,000
29
60
false
synth:6485255560750804075@1
synth
[{"aliases": ["src/binflow/core.py", "core.py"], "kind": "file", "value": "src/binflow/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/binflow/audit.py", "audit.py"], "kind": "file", "value": "src/binflow/audit.py"}, {"aliases": ["src/binf...
[{"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": ["bin_3.csv", "data/bin_3.csv"], "kind"...
ctx_compress-s1-00845
lotdesk
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 double counts carried amounts when the period closes on a weekend. It lives in src/lotdesk/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 lotdesk layout....
68,770
3,072
93
102
false
synth:2877710333198050000@1.1504
synth
[{"aliases": ["src/lotdesk/audit.py", "audit.py"], "kind": "file", "value": "src/lotdesk/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["data/stock_3.csv", "stock_3.csv"], "kind": "file", "value": "data/stock_3.csv"}, {"aliases": ["data/re...
[{"answers": ["src/lotdesk/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 most r...
ctx_compress-s1-00845-twin
lotdesk
ctx_compress-s1-00845
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 double counts carried amounts when the period closes on a weekend. It lives in src/lotdesk/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 lotdesk layout....
40,439
2,021
57
102
false
synth:2877710333198050000@1.1504
synth
[{"aliases": ["src/lotdesk/audit.py", "audit.py"], "kind": "file", "value": "src/lotdesk/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["data/stock_3.csv", "stock_3.csv"], "kind": "file", "value": "data/stock_3.csv"}, {"aliases": ["data/re...
[{"answers": ["src/lotdesk/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 most r...
ctx_compress-s1-00846
revisionworks
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] collate_heading_all averages one value too many once the window is full. It lives in src/revisionworks/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 revisionworks layou...
69,701
3,072
101
120
false
synth:252338209482192000@1.7778
synth
[{"aliases": ["src/revisionworks/cli.py", "cli.py"], "kind": "file", "value": "src/revisionworks/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["src/revisionworks/audit.py", "audit.py"], "kind": "file", "value": "src/revisionworks/audit.py"}, {"al...
[{"answers": ["src/revisionworks/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...
ctx_compress-s1-00846-twin
revisionworks
ctx_compress-s1-00846
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] collate_heading_all averages one value too many once the window is full. It lives in src/revisionworks/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 revisionworks layou...
34,243
1,712
61
120
false
synth:252338209482192000@1.7778
synth
[{"aliases": ["src/revisionworks/cli.py", "cli.py"], "kind": "file", "value": "src/revisionworks/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["src/revisionworks/audit.py", "audit.py"], "kind": "file", "value": "src/revisionworks/audit.py"}, {"al...
[{"answers": ["src/revisionworks/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...
ctx_compress-s1-00847
grantforge
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_policy produces doubled dashes for consecutive separators. It lives in src/grantforge/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 grantforge layout. [tool_...
41,316
2,065
65
76
false
synth:849675602187238260@1
synth
[{"aliases": ["src/grantforge/store.py", "store.py"], "kind": "file", "value": "src/grantforge/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/grantforge/compat.py", "compat.py"], "kind": "file", "value": "src/grantforge/compat.py"}, {...
[{"answers": ["src/grantforge/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-s1-00847-twin
grantforge
ctx_compress-s1-00847
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_policy produces doubled dashes for consecutive separators. It lives in src/grantforge/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 grantforge layout. [tool_...
27,053
1,352
45
76
false
synth:849675602187238260@1
synth
[{"aliases": ["src/grantforge/store.py", "store.py"], "kind": "file", "value": "src/grantforge/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/grantforge/compat.py", "compat.py"], "kind": "file", "value": "src/grantforge/compat.py"}, {...
[{"answers": ["src/grantforge/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-s1-00848
traceforge
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] clip_bucket rejects a range where lo equals hi, which is a valid single point range. It lives in src/traceforge/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 tracef...
72,469
3,072
95
118
false
synth:6453760059929175076@1.60712
synth
[{"aliases": ["src/traceforge/audit.py", "audit.py"], "kind": "file", "value": "src/traceforge/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["data/trace_7.csv", "trace_7.csv"], "kind": "file", "value": "data/trace_7.csv"}, {"aliases": ["s...
[{"answers": ["src/traceforge/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-s1-00848-twin
traceforge
ctx_compress-s1-00848
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] clip_bucket rejects a range where lo equals hi, which is a valid single point range. It lives in src/traceforge/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 tracef...
39,523
1,976
65
118
false
synth:6453760059929175076@1.60712
synth
[{"aliases": ["src/traceforge/audit.py", "audit.py"], "kind": "file", "value": "src/traceforge/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["data/trace_7.csv", "trace_7.csv"], "kind": "file", "value": "data/trace_7.csv"}, {"aliases": ["s...
[{"answers": ["src/traceforge/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-s1-00849
grantforge
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_session averages one value too many once the window is full. It lives in src/grantforge/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 grantforge layout. [tool_use ...
28,260
1,413
56
71
false
synth:5821640320517507749@1
synth
[{"aliases": ["src/grantforge/io.py", "io.py"], "kind": "file", "value": "src/grantforge/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["src/grantforge/report.py", "report.py"], "kind": "file", "value": "src/grantforge/report.py"}, {"aliases": ["src/g...
[{"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/grantforge/cli.py"], "k...
ctx_compress-s1-00849-twin
grantforge
ctx_compress-s1-00849
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_session averages one value too many once the window is full. It lives in src/grantforge/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 grantforge layout. [tool_use ...
17,827
1,000
38
71
false
synth:5821640320517507749@1
synth
[{"aliases": ["src/grantforge/io.py", "io.py"], "kind": "file", "value": "src/grantforge/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["src/grantforge/report.py", "report.py"], "kind": "file", "value": "src/grantforge/report.py"}, {"aliases": ["src/g...
[{"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_store.py", "tests/test_store.py"...
ctx_compress-s1-00850
bearingworks
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_waypoint_strict rejects a range where lo equals hi, which is a valid single point range. It lives in src/bearingworks/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 mappi...
66,476
3,072
89
114
false
synth:1069099584072745372@2.28003
synth
[{"aliases": ["src/bearingworks/format.py", "format.py"], "kind": "file", "value": "src/bearingworks/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["src/bearingworks/tables.py", "tables.py"], "kind": "file", "value": "src/bearingworks/...
[{"answers": ["src/bearingworks/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-s1-00850-twin
bearingworks
ctx_compress-s1-00850
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_waypoint_strict rejects a range where lo equals hi, which is a valid single point range. It lives in src/bearingworks/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 mappi...
42,095
2,104
57
114
false
synth:1069099584072745372@2.28003
synth
[{"aliases": ["src/bearingworks/format.py", "format.py"], "kind": "file", "value": "src/bearingworks/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["src/bearingworks/tables.py", "tables.py"], "kind": "file", "value": "src/bearingworks/...
[{"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": ["format.py", "src/bearingworks/format....
ctx_compress-s1-00851
sessionledger
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 skips the first page; page 1 returns what should be page 2. It lives in src/sessionledger/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 sessionledger layou...
58,421
2,921
87
106
false
synth:7649402878081978159@1.1469
synth
[{"aliases": ["src/sessionledger/cli.py", "cli.py"], "kind": "file", "value": "src/sessionledger/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["data/grant_2.csv", "grant_2.csv"], "kind": "file", "value": "data/grant_2.csv"}, {"aliases": ["docs/ch...
[{"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/sessionledger/compat...
ctx_compress-s1-00851-twin
sessionledger
ctx_compress-s1-00851
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 skips the first page; page 1 returns what should be page 2. It lives in src/sessionledger/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 sessionledger layou...
37,391
1,869
57
106
false
synth:7649402878081978159@1.1469
synth
[{"aliases": ["src/sessionledger/cli.py", "cli.py"], "kind": "file", "value": "src/sessionledger/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["data/grant_2.csv", "grant_2.csv"], "kind": "file", "value": "data/grant_2.csv"}, {"aliases": ["docs/ch...
[{"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-s1-00852
windowledger
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_window raises ZeroDivisionError on the first value and lags by one afterwards. It lives in src/windowledger/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 wi...
71,484
3,072
109
120
false
synth:2963676042429500933@1.42255
synth
[{"aliases": ["src/windowledger/rules.py", "rules.py"], "kind": "file", "value": "src/windowledger/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/windowledger/report.py", "report.py"], "kind": "file", "value": "src/windowledger/report...
[{"answers": ["src/windowledger/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-s1-00852-twin
windowledger
ctx_compress-s1-00852
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_window raises ZeroDivisionError on the first value and lags by one afterwards. It lives in src/windowledger/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 wi...
35,274
1,763
55
120
false
synth:2963676042429500933@1.42255
synth
[{"aliases": ["src/windowledger/rules.py", "rules.py"], "kind": "file", "value": "src/windowledger/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/windowledger/report.py", "report.py"], "kind": "file", "value": "src/windowledger/report...
[{"answers": ["src/windowledger/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-s1-00853
refundkit
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] settle_credit_safe skips the first page; page 1 returns what should be page 2. It lives in src/refundkit/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 refundkit l...
73,002
3,072
110
119
false
synth:7054030517484104707@1.23725
synth
[{"aliases": ["src/refundkit/format.py", "format.py"], "kind": "file", "value": "src/refundkit/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["src/refundkit/compat.py", "compat.py"], "kind": "file", "value": "src/refundkit/compat.py"},...
[{"answers": ["src/refundkit/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 mo...
ctx_compress-s1-00853-twin
refundkit
ctx_compress-s1-00853
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] settle_credit_safe skips the first page; page 1 returns what should be page 2. It lives in src/refundkit/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 refundkit l...
33,018
1,650
56
119
false
synth:7054030517484104707@1.23725
synth
[{"aliases": ["src/refundkit/format.py", "format.py"], "kind": "file", "value": "src/refundkit/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["src/refundkit/compat.py", "compat.py"], "kind": "file", "value": "src/refundkit/compat.py"},...
[{"answers": ["src/refundkit/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 mo...
ctx_compress-s1-00854
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] truncate_lexeme_strict returns one item too few on every page. It lives in src/lexemeledger/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 lexemeledger layout. [tool...
25,491
1,274
51
56
false
synth:7751097951542888090@1
synth
[{"aliases": ["src/lexemeledger/audit.py", "audit.py"], "kind": "file", "value": "src/lexemeledger/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["src/lexemeledger/compat.py", "compat.py"], "kind": "file", "value": "src/lexemeledger/compat...
[{"answers": ["src/lexemeledger/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 m...
ctx_compress-s1-00854-twin
lexemeledger
ctx_compress-s1-00854
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_lexeme_strict returns one item too few on every page. It lives in src/lexemeledger/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 lexemeledger layout. [tool...
16,990
1,000
27
56
false
synth:7751097951542888090@1
synth
[{"aliases": ["src/lexemeledger/audit.py", "audit.py"], "kind": "file", "value": "src/lexemeledger/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["src/lexemeledger/compat.py", "compat.py"], "kind": "file", "value": "src/lexemeledger/compat...
[{"answers": ["src/lexemeledger/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 m...