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-s1-01256-twin
stockdesk
ctx_compress-s1-01256
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] age_warehouse raises SkuBoundError for bytes input in strict mode instead of decoding it. It lives in src/stockdesk/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 st...
11,115
1,000
27
54
false
synth:8967116792875749760@1
synth
[{"aliases": ["src/stockdesk/store.py", "store.py"], "kind": "file", "value": "src/stockdesk/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/stockdesk/tables.py", "tables.py"], "kind": "file", "value": "src/stockdesk/tables.py"}, {"ali...
[{"answers": ["src/stockdesk/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...
ctx_compress-s1-01257
tariffflow
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/tariffflow/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 tariffflow layout. [t...
100,277
3,072
126
155
false
synth:2854943770961551854@1.83435
synth
[{"aliases": ["src/tariffflow/compat.py", "compat.py"], "kind": "file", "value": "src/tariffflow/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["data/refund_11.csv", "refund_11.csv"], "kind": "file", "value": "data/refund_11.csv"}, {"a...
[{"answers": ["src/tariffflow/compat.py", "compat.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's m...
ctx_compress-s1-01257-twin
tariffflow
ctx_compress-s1-01257
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/tariffflow/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 tariffflow layout. [t...
56,737
2,836
76
155
false
synth:2854943770961551854@1.83435
synth
[{"aliases": ["src/tariffflow/compat.py", "compat.py"], "kind": "file", "value": "src/tariffflow/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["data/refund_11.csv", "refund_11.csv"], "kind": "file", "value": "data/refund_11.csv"}, {"a...
[{"answers": ["src/tariffflow/compat.py", "compat.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's m...
ctx_compress-s1-01258
binkit
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] pick_pallet_once drops override values of 0, False and empty string. It lives in src/binkit/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 binkit layout. [tool_use k...
68,245
3,072
112
121
false
synth:7869524925046815782@2.34088
synth
[{"aliases": ["src/binkit/store.py", "store.py"], "kind": "file", "value": "src/binkit/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["data/reorder_0.csv", "reorder_0.csv"], "kind": "file", "value": "data/reorder_0.csv"}, {"aliases": ["src...
[{"answers": ["src/binkit/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 re...
ctx_compress-s1-01258-twin
binkit
ctx_compress-s1-01258
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_pallet_once drops override values of 0, False and empty string. It lives in src/binkit/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 binkit layout. [tool_use k...
36,083
1,804
61
121
false
synth:7869524925046815782@2.34088
synth
[{"aliases": ["src/binkit/store.py", "store.py"], "kind": "file", "value": "src/binkit/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["data/reorder_0.csv", "reorder_0.csv"], "kind": "file", "value": "data/reorder_0.csv"}, {"aliases": ["src...
[{"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_limits.py", "tests/test_limits.p...
ctx_compress-s1-01259
seriesdesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] merge_gauge_strict double counts carried amounts when the period closes on a weekend. It lives in src/seriesdesk/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 seriesd...
106,049
3,072
135
148
false
synth:1612313367675195430@1.94475
synth
[{"aliases": ["src/seriesdesk/core.py", "core.py"], "kind": "file", "value": "src/seriesdesk/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["data/probe_11.csv", "probe_11.csv"], "kind": "file", "value": "data/probe_11.csv"}, {"aliases": ["src/...
[{"answers": ["src/seriesdesk/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-s1-01259-twin
seriesdesk
ctx_compress-s1-01259
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] merge_gauge_strict double counts carried amounts when the period closes on a weekend. It lives in src/seriesdesk/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 seriesd...
42,146
2,107
73
148
false
synth:1612313367675195430@1.94475
synth
[{"aliases": ["src/seriesdesk/core.py", "core.py"], "kind": "file", "value": "src/seriesdesk/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["data/probe_11.csv", "probe_11.csv"], "kind": "file", "value": "data/probe_11.csv"}, {"aliases": ["src/...
[{"answers": ["src/seriesdesk/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-s1-01260
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] pick_reorder_safe keeps leading spaces in keys so lookups after a semicolon fail. It lives in src/binflow/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 binflow layo...
20,495
1,024
58
73
false
synth:6853983281039324150@1
synth
[{"aliases": ["src/binflow/rules.py", "rules.py"], "kind": "file", "value": "src/binflow/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/binflow/schema.py", "schema.py"], "kind": "file", "value": "src/binflow/schema.py"}, {"aliases": [...
[{"answers": ["src/binflow/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 r...
ctx_compress-s1-01260-twin
binflow
ctx_compress-s1-01260
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_reorder_safe keeps leading spaces in keys so lookups after a semicolon fail. It lives in src/binflow/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 binflow layo...
13,498
1,000
38
73
false
synth:6853983281039324150@1
synth
[{"aliases": ["src/binflow/rules.py", "rules.py"], "kind": "file", "value": "src/binflow/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/binflow/schema.py", "schema.py"], "kind": "file", "value": "src/binflow/schema.py"}, {"aliases": [...
[{"answers": ["src/binflow/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 r...
ctx_compress-s1-01261
counterdesk
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] merge_probe double counts carried amounts when the period closes on a weekend. It lives in src/counterdesk/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 counterdesk lay...
45,575
2,278
71
80
false
synth:5136056226629323062@1
synth
[{"aliases": ["src/counterdesk/cli.py", "cli.py"], "kind": "file", "value": "src/counterdesk/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["src/counterdesk/compat.py", "compat.py"], "kind": "file", "value": "src/counterdesk/compat.py"}, {"aliases...
[{"answers": ["src/counterdesk/cli.py", "cli.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most r...
ctx_compress-s1-01261-twin
counterdesk
ctx_compress-s1-01261
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] merge_probe double counts carried amounts when the period closes on a weekend. It lives in src/counterdesk/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 counterdesk lay...
20,510
1,025
37
80
false
synth:5136056226629323062@1
synth
[{"aliases": ["src/counterdesk/cli.py", "cli.py"], "kind": "file", "value": "src/counterdesk/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["src/counterdesk/compat.py", "compat.py"], "kind": "file", "value": "src/counterdesk/compat.py"}, {"aliases...
[{"answers": ["src/counterdesk/cli.py", "cli.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most r...
ctx_compress-s1-01262
lotledger
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_lot_batch raises ZeroDivisionError on the first value and lags by one afterwards. It lives in src/lotledger/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 lotl...
87,258
3,072
107
122
false
synth:6789960516694318011@1.8608
synth
[{"aliases": ["src/lotledger/audit.py", "audit.py"], "kind": "file", "value": "src/lotledger/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["src/lotledger/io.py", "io.py"], "kind": "file", "value": "src/lotledger/io.py"}, {"aliases": ["dat...
[{"answers": ["src/lotledger/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...
ctx_compress-s1-01262-twin
lotledger
ctx_compress-s1-01262
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_lot_batch raises ZeroDivisionError on the first value and lags by one afterwards. It lives in src/lotledger/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 lotl...
44,965
2,248
67
122
false
synth:6789960516694318011@1.8608
synth
[{"aliases": ["src/lotledger/audit.py", "audit.py"], "kind": "file", "value": "src/lotledger/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["src/lotledger/io.py", "io.py"], "kind": "file", "value": "src/lotledger/io.py"}, {"aliases": ["dat...
[{"answers": ["src/lotledger/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...
ctx_compress-s1-01263
tilekit
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_bearing skips the first page; page 1 returns what should be page 2. It lives in src/tilekit/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 tilekit layout. [tool_use k...
65,085
3,072
85
108
false
synth:8889623342287562387@1.682
synth
[{"aliases": ["src/tilekit/io.py", "io.py"], "kind": "file", "value": "src/tilekit/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["src/tilekit/core.py", "core.py"], "kind": "file", "value": "src/tilekit/core.py"}, {"aliases": ["data/bearing_1.csv", "b...
[{"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-s1-01263-twin
tilekit
ctx_compress-s1-01263
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_bearing skips the first page; page 1 returns what should be page 2. It lives in src/tilekit/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 tilekit layout. [tool_use k...
43,543
2,177
59
108
false
synth:8889623342287562387@1.682
synth
[{"aliases": ["src/tilekit/io.py", "io.py"], "kind": "file", "value": "src/tilekit/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["src/tilekit/core.py", "core.py"], "kind": "file", "value": "src/tilekit/core.py"}, {"aliases": ["data/bearing_1.csv", "b...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["core.py", "src/tilekit/core.py"], "ki...
ctx_compress-s1-01264
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] bin_bucket raises ProbeBoundError for bytes input in strict mode instead of decoding it. It lives in src/counterflow/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...
44,621
2,231
60
71
false
synth:6634779417921435993@1
synth
[{"aliases": ["src/counterflow/compat.py", "compat.py"], "kind": "file", "value": "src/counterflow/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"}, {"al...
[{"answers": ["src/counterflow/compat.py", "compat.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's ...
ctx_compress-s1-01264-twin
counterflow
ctx_compress-s1-01264
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_bucket raises ProbeBoundError for bytes input in strict mode instead of decoding it. It lives in src/counterflow/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...
26,325
1,316
36
71
false
synth:6634779417921435993@1
synth
[{"aliases": ["src/counterflow/compat.py", "compat.py"], "kind": "file", "value": "src/counterflow/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"}, {"al...
[{"answers": ["src/counterflow/compat.py", "compat.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's ...
ctx_compress-s1-01265
sessionflow
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_page misses keys that differ only in case. It lives in src/sessionflow/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 sessionflow layout. [tool_use kind=sea...
100,376
3,072
135
148
false
synth:5694969972700320945@2.2313
synth
[{"aliases": ["src/sessionflow/io.py", "io.py"], "kind": "file", "value": "src/sessionflow/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["data/nonce_8.csv", "nonce_8.csv"], "kind": "file", "value": "data/nonce_8.csv"}, {"aliases": ["src/sessionflow/c...
[{"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_limits.py", "tests/test_limits.p...
ctx_compress-s1-01265-twin
sessionflow
ctx_compress-s1-01265
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_page misses keys that differ only in case. It lives in src/sessionflow/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 sessionflow layout. [tool_use kind=sea...
61,407
3,070
85
148
false
synth:5694969972700320945@2.2313
synth
[{"aliases": ["src/sessionflow/io.py", "io.py"], "kind": "file", "value": "src/sessionflow/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["data/nonce_8.csv", "nonce_8.csv"], "kind": "file", "value": "data/nonce_8.csv"}, {"aliases": ["src/sessionflow/c...
[{"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_2.csv", "session_2.csv"]...
ctx_compress-s1-01266
probekit
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_probe_once ignores the cap, so late attempts wait for minutes. It lives in src/probekit/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 probekit layout. [tool_us...
30,440
1,522
64
69
false
synth:5301915531799721334@1
synth
[{"aliases": ["src/probekit/rules.py", "rules.py"], "kind": "file", "value": "src/probekit/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["data/span_0.csv", "span_0.csv"], "kind": "file", "value": "data/span_0.csv"}, {"aliases": ["src/prob...
[{"answers": ["src/probekit/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-s1-01266-twin
probekit
ctx_compress-s1-01266
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_probe_once ignores the cap, so late attempts wait for minutes. It lives in src/probekit/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 probekit layout. [tool_us...
17,334
1,000
38
69
false
synth:5301915531799721334@1
synth
[{"aliases": ["src/probekit/rules.py", "rules.py"], "kind": "file", "value": "src/probekit/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["data/span_0.csv", "span_0.csv"], "kind": "file", "value": "data/span_0.csv"}, {"aliases": ["src/prob...
[{"answers": ["src/probekit/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-s1-01267
marginkit
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] redact_footnote_once raises ZeroDivisionError on the first value and lags by one afterwards. It lives in src/marginkit/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...
85,308
3,072
95
112
false
synth:3546073950539045880@1.5026
synth
[{"aliases": ["src/marginkit/store.py", "store.py"], "kind": "file", "value": "src/marginkit/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["data/revision_2.csv", "revision_2.csv"], "kind": "file", "value": "data/revision_2.csv"}, {"aliase...
[{"answers": ["src/marginkit/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...
ctx_compress-s1-01267-twin
marginkit
ctx_compress-s1-01267
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] redact_footnote_once raises ZeroDivisionError on the first value and lags by one afterwards. It lives in src/marginkit/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...
46,471
2,323
59
112
false
synth:3546073950539045880@1.5026
synth
[{"aliases": ["src/marginkit/store.py", "store.py"], "kind": "file", "value": "src/marginkit/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["data/revision_2.csv", "revision_2.csv"], "kind": "file", "value": "data/revision_2.csv"}, {"aliase...
[{"answers": ["src/marginkit/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 most...
ctx_compress-s1-01268
footnoteledger
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_footnote keeps leading spaces in keys so lookups after a semicolon fail. It lives in src/footnoteledger/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 footnot...
41,520
2,076
81
96
false
synth:8023168155859360886@1.27168
synth
[{"aliases": ["src/footnoteledger/store.py", "store.py"], "kind": "file", "value": "src/footnoteledger/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["data/section_3.csv", "section_3.csv"], "kind": "file", "value": "data/section_3.csv"}, {...
[{"answers": ["src/footnoteledger/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...
ctx_compress-s1-01268-twin
footnoteledger
ctx_compress-s1-01268
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_footnote keeps leading spaces in keys so lookups after a semicolon fail. It lives in src/footnoteledger/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 footnot...
28,090
1,404
49
96
false
synth:8023168155859360886@1.27168
synth
[{"aliases": ["src/footnoteledger/store.py", "store.py"], "kind": "file", "value": "src/footnoteledger/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["data/section_3.csv", "section_3.csv"], "kind": "file", "value": "data/section_3.csv"}, {...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["io.py", "src/footnoteledger/io.py"], ...
ctx_compress-s1-01269
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] reserve_bin_batch raises BinStateError for bytes input in strict mode instead of decoding it. It lives in src/palletkit/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...
43,549
2,177
70
81
false
synth:799816697877400781@1
synth
[{"aliases": ["src/palletkit/rules.py", "rules.py"], "kind": "file", "value": "src/palletkit/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["data/sku_3.csv", "sku_3.csv"], "kind": "file", "value": "data/sku_3.csv"}, {"aliases": ["src/palle...
[{"answers": ["src/palletkit/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-s1-01269-twin
palletkit
ctx_compress-s1-01269
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_bin_batch raises BinStateError for bytes input in strict mode instead of decoding it. It lives in src/palletkit/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...
31,100
1,555
48
81
false
synth:799816697877400781@1
synth
[{"aliases": ["src/palletkit/rules.py", "rules.py"], "kind": "file", "value": "src/palletkit/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["data/sku_3.csv", "sku_3.csv"], "kind": "file", "value": "data/sku_3.csv"}, {"aliases": ["src/palle...
[{"answers": ["src/palletkit/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-s1-01270
fixworks
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] project_beacon_strict lets defaults overwrite explicit overrides. It lives in src/fixworks/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 fixworks layout. [tool_use ...
69,274
3,072
95
106
false
synth:7770449794805863768@1.14685
synth
[{"aliases": ["src/fixworks/rules.py", "rules.py"], "kind": "file", "value": "src/fixworks/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/fixworks/store.py", "store.py"], "kind": "file", "value": "src/fixworks/store.py"}, {"aliases": ...
[{"answers": ["src/fixworks/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-s1-01270-twin
fixworks
ctx_compress-s1-01270
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] project_beacon_strict lets defaults overwrite explicit overrides. It lives in src/fixworks/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 fixworks layout. [tool_use ...
32,772
1,638
53
106
false
synth:7770449794805863768@1.14685
synth
[{"aliases": ["src/fixworks/rules.py", "rules.py"], "kind": "file", "value": "src/fixworks/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/fixworks/store.py", "store.py"], "kind": "file", "value": "src/fixworks/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": ["rules.py", "src/fixworks/rules.py"], ...
ctx_compress-s1-01271
grantdesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] rotate_scope_window grows linearly instead of doubling. It lives in src/grantdesk/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 grantdesk layout. [tool_use kind=s...
56,423
2,821
84
107
false
synth:1184153788654661365@1.1295
synth
[{"aliases": ["src/grantdesk/compat.py", "compat.py"], "kind": "file", "value": "src/grantdesk/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["src/grantdesk/rules.py", "rules.py"], "kind": "file", "value": "src/grantdesk/rules.py"}, {"...
[{"answers": ["src/grantdesk/compat.py", "compat.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's mo...
ctx_compress-s1-01271-twin
grantdesk
ctx_compress-s1-01271
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] rotate_scope_window grows linearly instead of doubling. It lives in src/grantdesk/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 grantdesk layout. [tool_use kind=s...
39,822
1,991
62
107
false
synth:1184153788654661365@1.1295
synth
[{"aliases": ["src/grantdesk/compat.py", "compat.py"], "kind": "file", "value": "src/grantdesk/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["src/grantdesk/rules.py", "rules.py"], "kind": "file", "value": "src/grantdesk/rules.py"}, {"...
[{"answers": ["src/grantdesk/compat.py", "compat.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's mo...
ctx_compress-s1-01272
marginflow
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_once misses keys that differ only in case. It lives in src/marginflow/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 marginflow layout. [tool_use kin...
25,109
1,255
54
63
false
synth:468791315404557623@1
synth
[{"aliases": ["src/marginflow/store.py", "store.py"], "kind": "file", "value": "src/marginflow/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/marginflow/tables.py", "tables.py"], "kind": "file", "value": "src/marginflow/tables.py"}, {...
[{"answers": ["src/marginflow/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-s1-01272-twin
marginflow
ctx_compress-s1-01272
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_once misses keys that differ only in case. It lives in src/marginflow/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 marginflow layout. [tool_use kin...
14,497
1,000
32
63
false
synth:468791315404557623@1
synth
[{"aliases": ["src/marginflow/store.py", "store.py"], "kind": "file", "value": "src/marginflow/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/marginflow/tables.py", "tables.py"], "kind": "file", "value": "src/marginflow/tables.py"}, {...
[{"answers": ["src/marginflow/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-s1-01273
stockflow
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] restock_lot grows linearly instead of doubling. It lives in src/stockflow/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 stockflow layout. [tool_use kind=search name=B...
46,043
2,302
77
94
false
synth:7685936703119783249@1
synth
[{"aliases": ["src/stockflow/core.py", "core.py"], "kind": "file", "value": "src/stockflow/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["data/pallet_2.csv", "pallet_2.csv"], "kind": "file", "value": "data/pallet_2.csv"}, {"aliases": ["src/st...
[{"answers": ["src/stockflow/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-s1-01273-twin
stockflow
ctx_compress-s1-01273
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] restock_lot grows linearly instead of doubling. It lives in src/stockflow/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 stockflow layout. [tool_use kind=search name=B...
21,921
1,096
43
94
false
synth:7685936703119783249@1
synth
[{"aliases": ["src/stockflow/core.py", "core.py"], "kind": "file", "value": "src/stockflow/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["data/pallet_2.csv", "pallet_2.csv"], "kind": "file", "value": "data/pallet_2.csv"}, {"aliases": ["src/st...
[{"answers": ["src/stockflow/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-s1-01274
glyphkit
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] tokenise_lexeme returns one item too few on every page. It lives in src/glyphkit/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 glyphkit layout. [tool_use kind=sea...
83,714
3,072
99
112
false
synth:7586406698485464120@1.35312
synth
[{"aliases": ["src/glyphkit/compat.py", "compat.py"], "kind": "file", "value": "src/glyphkit/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["data/slug_4.csv", "slug_4.csv"], "kind": "file", "value": "data/slug_4.csv"}, {"aliases": ["da...
[{"answers": ["src/glyphkit/compat.py", "compat.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's mos...
ctx_compress-s1-01274-twin
glyphkit
ctx_compress-s1-01274
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] tokenise_lexeme returns one item too few on every page. It lives in src/glyphkit/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 glyphkit layout. [tool_use kind=sea...
51,609
2,580
63
112
false
synth:7586406698485464120@1.35312
synth
[{"aliases": ["src/glyphkit/compat.py", "compat.py"], "kind": "file", "value": "src/glyphkit/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["data/slug_4.csv", "slug_4.csv"], "kind": "file", "value": "data/slug_4.csv"}, {"aliases": ["da...
[{"answers": ["src/glyphkit/compat.py", "compat.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's mos...
ctx_compress-s1-01275
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] extend_calendar ignores the cap, so late attempts wait for minutes. It lives in src/shiftflow/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 shiftflow layout. [tool_...
60,322
3,016
103
122
false
synth:2916879132401886818@1.08
synth
[{"aliases": ["src/shiftflow/store.py", "store.py"], "kind": "file", "value": "src/shiftflow/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/shiftflow/tables.py", "tables.py"], "kind": "file", "value": "src/shiftflow/tables.py"}, {"ali...
[{"answers": ["src/shiftflow/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...
ctx_compress-s1-01275-twin
shiftflow
ctx_compress-s1-01275
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 ignores the cap, so late attempts wait for minutes. It lives in src/shiftflow/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 shiftflow layout. [tool_...
32,471
1,623
63
122
false
synth:2916879132401886818@1.08
synth
[{"aliases": ["src/shiftflow/store.py", "store.py"], "kind": "file", "value": "src/shiftflow/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/shiftflow/tables.py", "tables.py"], "kind": "file", "value": "src/shiftflow/tables.py"}, {"ali...
[{"answers": ["src/shiftflow/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...
ctx_compress-s1-01276
ledgerforge
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] accrue_payout returns the upper bound for every value instead of clamping into the range. It lives in src/ledgerforge/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 th...
61,679
3,072
90
111
false
synth:2332905891799339423@1.99832
synth
[{"aliases": ["src/ledgerforge/format.py", "format.py"], "kind": "file", "value": "src/ledgerforge/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["src/ledgerforge/audit.py", "audit.py"], "kind": "file", "value": "src/ledgerforge/audit....
[{"answers": ["src/ledgerforge/format.py", "format.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's ...
ctx_compress-s1-01276-twin
ledgerforge
ctx_compress-s1-01276
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_payout returns the upper bound for every value instead of clamping into the range. It lives in src/ledgerforge/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 th...
40,068
2,003
63
111
false
synth:2332905891799339423@1.99832
synth
[{"aliases": ["src/ledgerforge/format.py", "format.py"], "kind": "file", "value": "src/ledgerforge/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["src/ledgerforge/audit.py", "audit.py"], "kind": "file", "value": "src/ledgerforge/audit....
[{"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/ledgerforge/format.p...
ctx_compress-s1-01277
payoutkit
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] allocate_receipt_safe drops override values of 0, False and empty string. It lives in src/payoutkit/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 payoutkit layout...
83,539
3,072
104
131
false
synth:5874607654279328702@1.98075
synth
[{"aliases": ["src/payoutkit/compat.py", "compat.py"], "kind": "file", "value": "src/payoutkit/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"}, {"aliase...
[{"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": ["balance_1.csv", "data/balance_1.csv"]...
ctx_compress-s1-01277-twin
payoutkit
ctx_compress-s1-01277
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] allocate_receipt_safe drops override values of 0, False and empty string. It lives in src/payoutkit/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 payoutkit layout...
50,903
2,545
70
131
false
synth:5874607654279328702@1.98075
synth
[{"aliases": ["src/payoutkit/compat.py", "compat.py"], "kind": "file", "value": "src/payoutkit/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"}, {"aliase...
[{"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-s1-01278
principalflow
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_scope raises PrincipalRangeError for bytes input in strict mode instead of decoding it. It lives in src/principalflow/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 mappi...
100,926
3,072
114
143
false
synth:6017556336575901524@1.87157
synth
[{"aliases": ["src/principalflow/audit.py", "audit.py"], "kind": "file", "value": "src/principalflow/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["src/principalflow/io.py", "io.py"], "kind": "file", "value": "src/principalflow/io.py"}, {...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["data/policy_4.csv", "policy_4.csv"], ...
ctx_compress-s1-01278-twin
principalflow
ctx_compress-s1-01278
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_scope raises PrincipalRangeError for bytes input in strict mode instead of decoding it. It lives in src/principalflow/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 mappi...
62,307
3,072
78
143
false
synth:6017556336575901524@1.87157
synth
[{"aliases": ["src/principalflow/audit.py", "audit.py"], "kind": "file", "value": "src/principalflow/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["src/principalflow/io.py", "io.py"], "kind": "file", "value": "src/principalflow/io.py"}, {...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["data/grant_10.csv", "grant_10.csv"], ...
ctx_compress-s1-01279
policyflow
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] narrow_principal misses keys that differ only in case. It lives in src/policyflow/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 policyflow layout. [tool_use kind=se...
57,544
2,877
86
105
false
synth:3469024119572357246@1.42568
synth
[{"aliases": ["src/policyflow/store.py", "store.py"], "kind": "file", "value": "src/policyflow/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/policyflow/schema.py", "schema.py"], "kind": "file", "value": "src/policyflow/schema.py"}, {...
[{"answers": ["src/policyflow/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-s1-01279-twin
policyflow
ctx_compress-s1-01279
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] narrow_principal misses keys that differ only in case. It lives in src/policyflow/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 policyflow layout. [tool_use kind=se...
31,381
1,569
49
105
false
synth:3469024119572357246@1.42568
synth
[{"aliases": ["src/policyflow/store.py", "store.py"], "kind": "file", "value": "src/policyflow/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/policyflow/schema.py", "schema.py"], "kind": "file", "value": "src/policyflow/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": ["schema.py", "src/policyflow/schema.py...
ctx_compress-s1-01280
glyphdesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] wrap_cue rejects a range where lo equals hi, which is a valid single point range. It lives in src/glyphdesk/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 glyphdesk layo...
92,527
3,072
112
121
false
synth:6099661903887509742@1.76228
synth
[{"aliases": ["src/glyphdesk/cli.py", "cli.py"], "kind": "file", "value": "src/glyphdesk/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["src/glyphdesk/audit.py", "audit.py"], "kind": "file", "value": "src/glyphdesk/audit.py"}, {"aliases": ["data/l...
[{"answers": ["src/glyphdesk/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 rec...
ctx_compress-s1-01280-twin
glyphdesk
ctx_compress-s1-01280
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] wrap_cue rejects a range where lo equals hi, which is a valid single point range. It lives in src/glyphdesk/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 glyphdesk layo...
43,428
2,171
60
121
false
synth:6099661903887509742@1.76228
synth
[{"aliases": ["src/glyphdesk/cli.py", "cli.py"], "kind": "file", "value": "src/glyphdesk/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["src/glyphdesk/audit.py", "audit.py"], "kind": "file", "value": "src/glyphdesk/audit.py"}, {"aliases": ["data/l...
[{"answers": ["src/glyphdesk/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 rec...
ctx_compress-s1-01281
principalforge
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_policy raises KeyError for a missing key instead of returning the default. It lives in src/principalforge/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 pri...
51,837
2,591
88
95
false
synth:3205350753911620376@1
synth
[{"aliases": ["src/principalforge/limits.py", "limits.py"], "kind": "file", "value": "src/principalforge/limits.py"}, {"aliases": ["tests/test_limits.py", "test_limits.py"], "kind": "file", "value": "tests/test_limits.py"}, {"aliases": ["data/session_2.csv", "session_2.csv"], "kind": "file", "value": "data/session_2.cs...
[{"answers": ["src/principalforge/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...
ctx_compress-s1-01281-twin
principalforge
ctx_compress-s1-01281
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_policy raises KeyError for a missing key instead of returning the default. It lives in src/principalforge/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 pri...
28,665
1,433
48
95
false
synth:3205350753911620376@1
synth
[{"aliases": ["src/principalforge/limits.py", "limits.py"], "kind": "file", "value": "src/principalforge/limits.py"}, {"aliases": ["tests/test_limits.py", "test_limits.py"], "kind": "file", "value": "tests/test_limits.py"}, {"aliases": ["data/session_2.csv", "session_2.csv"], "kind": "file", "value": "data/session_2.cs...
[{"answers": ["src/principalforge/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...
ctx_compress-s1-01282
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] net_payout rejects a range where lo equals hi, which is a valid single point range. It lives in src/tariffkit/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 tariffki...
66,908
3,072
92
111
false
synth:470638059098264224@1.5339
synth
[{"aliases": ["src/tariffkit/rules.py", "rules.py"], "kind": "file", "value": "src/tariffkit/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/tariffkit/report.py", "report.py"], "kind": "file", "value": "src/tariffkit/report.py"}, {"ali...
[{"answers": ["src/tariffkit/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-s1-01282-twin
tariffkit
ctx_compress-s1-01282
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_payout rejects a range where lo equals hi, which is a valid single point range. It lives in src/tariffkit/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 tariffki...
41,383
2,069
62
111
false
synth:470638059098264224@1.5339
synth
[{"aliases": ["src/tariffkit/rules.py", "rules.py"], "kind": "file", "value": "src/tariffkit/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/tariffkit/report.py", "report.py"], "kind": "file", "value": "src/tariffkit/report.py"}, {"ali...
[{"answers": ["src/tariffkit/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-s1-01283
bookingkit
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] assign_shift_all skips the first page; page 1 returns what should be page 2. It lives in src/bookingkit/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 bookingkit l...
61,097
3,054
102
113
false
synth:2945155152659572498@1.32088
synth
[{"aliases": ["src/bookingkit/report.py", "report.py"], "kind": "file", "value": "src/bookingkit/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["data/shift_4.csv", "shift_4.csv"], "kind": "file", "value": "data/shift_4.csv"}, {"aliases...
[{"answers": ["src/bookingkit/report.py", "report.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's m...
ctx_compress-s1-01283-twin
bookingkit
ctx_compress-s1-01283
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] assign_shift_all skips the first page; page 1 returns what should be page 2. It lives in src/bookingkit/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 bookingkit l...
32,656
1,632
62
113
false
synth:2945155152659572498@1.32088
synth
[{"aliases": ["src/bookingkit/report.py", "report.py"], "kind": "file", "value": "src/bookingkit/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["data/shift_4.csv", "shift_4.csv"], "kind": "file", "value": "data/shift_4.csv"}, {"aliases...
[{"answers": ["src/bookingkit/report.py", "report.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's m...
ctx_compress-s1-01284
ledgerledger
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_ledger_safe returns one item too few on every page. It lives in src/ledgerledger/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 ledgerledger layout. [tool_us...
54,308
2,715
104
111
false
synth:1919612510164913273@1.13665
synth
[{"aliases": ["src/ledgerledger/core.py", "core.py"], "kind": "file", "value": "src/ledgerledger/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["data/payout_1.csv", "payout_1.csv"], "kind": "file", "value": "data/payout_1.csv"}, {"aliases": ["...
[{"answers": ["src/ledgerledger/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-s1-01284-twin
ledgerledger
ctx_compress-s1-01284
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_ledger_safe returns one item too few on every page. It lives in src/ledgerledger/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 ledgerledger layout. [tool_us...
35,021
1,751
66
111
false
synth:1919612510164913273@1.13665
synth
[{"aliases": ["src/ledgerledger/core.py", "core.py"], "kind": "file", "value": "src/ledgerledger/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["data/payout_1.csv", "payout_1.csv"], "kind": "file", "value": "data/payout_1.csv"}, {"aliases": ["...
[{"answers": ["src/ledgerledger/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-s1-01285
snippetworks
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] wrap_snippet raises ChunkShapeError for bytes input in strict mode instead of decoding it. It lives in src/snippetworks/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 ...
25,662
1,283
61
72
false
synth:4437188414262989197@1
synth
[{"aliases": ["src/snippetworks/limits.py", "limits.py"], "kind": "file", "value": "src/snippetworks/limits.py"}, {"aliases": ["tests/test_limits.py", "test_limits.py"], "kind": "file", "value": "tests/test_limits.py"}, {"aliases": ["data/glyph_4.csv", "glyph_4.csv"], "kind": "file", "value": "data/glyph_4.csv"}, {"ali...
[{"answers": ["src/snippetworks/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-s1-01285-twin
snippetworks
ctx_compress-s1-01285
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] wrap_snippet raises ChunkShapeError for bytes input in strict mode instead of decoding it. It lives in src/snippetworks/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 ...
12,317
1,000
35
72
false
synth:4437188414262989197@1
synth
[{"aliases": ["src/snippetworks/limits.py", "limits.py"], "kind": "file", "value": "src/snippetworks/limits.py"}, {"aliases": ["tests/test_limits.py", "test_limits.py"], "kind": "file", "value": "tests/test_limits.py"}, {"aliases": ["data/glyph_4.csv", "glyph_4.csv"], "kind": "file", "value": "data/glyph_4.csv"}, {"ali...
[{"answers": ["src/snippetworks/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-s1-01286
snippetdesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] wrap_stem produces doubled dashes for consecutive separators. It lives in src/snippetdesk/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 snippetdesk layout. [tool_use ...
49,359
2,467
86
101
false
synth:7651953119128081370@1.47073
synth
[{"aliases": ["src/snippetdesk/core.py", "core.py"], "kind": "file", "value": "src/snippetdesk/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["data/glyph_1.csv", "glyph_1.csv"], "kind": "file", "value": "data/glyph_1.csv"}, {"aliases": ["src/s...
[{"answers": ["src/snippetdesk/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-s1-01286-twin
snippetdesk
ctx_compress-s1-01286
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] wrap_stem produces doubled dashes for consecutive separators. It lives in src/snippetdesk/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 snippetdesk layout. [tool_use ...
27,093
1,354
49
101
false
synth:7651953119128081370@1.47073
synth
[{"aliases": ["src/snippetdesk/core.py", "core.py"], "kind": "file", "value": "src/snippetdesk/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["data/glyph_1.csv", "glyph_1.csv"], "kind": "file", "value": "data/glyph_1.csv"}, {"aliases": ["src/s...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["data/glyph_1.csv", "glyph_1.csv"], "k...
ctx_compress-s1-01287
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] settle_refund_once raises ZeroDivisionError on the first value and lags by one afterwards. 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 invoi...
83,434
3,072
102
125
false
synth:8178398868375527118@1.72867
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": ["data/tariff_0.csv", "tariff_0.csv"], "kind": "file", "value": "data/tariff_0.csv"}, {"aliases": ["src/invoicedes...
[{"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_limits.py", "tests/test_limits.p...
ctx_compress-s1-01287-twin
invoicedesk
ctx_compress-s1-01287
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_refund_once raises ZeroDivisionError on the first value and lags by one afterwards. 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 invoi...
58,257
2,912
74
125
false
synth:8178398868375527118@1.72867
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": ["data/tariff_0.csv", "tariff_0.csv"], "kind": "file", "value": "data/tariff_0.csv"}, {"aliases": ["src/invoicedes...
[{"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/invoicedesk/tables.py", "tables.p...
ctx_compress-s1-01288
traceworks
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_span_window keeps leading spaces in keys so lookups after a semicolon fail. It lives in src/traceworks/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 tracew...
87,635
3,072
120
147
false
synth:7358745035875793025@2.35712
synth
[{"aliases": ["src/traceworks/report.py", "report.py"], "kind": "file", "value": "src/traceworks/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["src/traceworks/cli.py", "cli.py"], "kind": "file", "value": "src/traceworks/cli.py"}, {"al...
[{"answers": ["src/traceworks/report.py", "report.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's m...
ctx_compress-s1-01288-twin
traceworks
ctx_compress-s1-01288
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_span_window keeps leading spaces in keys so lookups after a semicolon fail. It lives in src/traceworks/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 tracew...
49,548
2,477
71
147
false
synth:7358745035875793025@2.35712
synth
[{"aliases": ["src/traceworks/report.py", "report.py"], "kind": "file", "value": "src/traceworks/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["src/traceworks/cli.py", "cli.py"], "kind": "file", "value": "src/traceworks/cli.py"}, {"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": ["test_cli.py", "tests/test_cli.py"], "...
ctx_compress-s1-01289
stemkit
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] truncate_slug_strict produces doubled dashes for consecutive separators. It lives in src/stemkit/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 stemkit layout. [tool...
55,034
2,751
85
104
false
synth:4055590786500085328@1.13247
synth
[{"aliases": ["src/stemkit/audit.py", "audit.py"], "kind": "file", "value": "src/stemkit/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["src/stemkit/tables.py", "tables.py"], "kind": "file", "value": "src/stemkit/tables.py"}, {"aliases": [...
[{"answers": ["src/stemkit/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-s1-01289-twin
stemkit
ctx_compress-s1-01289
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_slug_strict produces doubled dashes for consecutive separators. It lives in src/stemkit/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 stemkit layout. [tool...
31,253
1,562
61
104
false
synth:4055590786500085328@1.13247
synth
[{"aliases": ["src/stemkit/audit.py", "audit.py"], "kind": "file", "value": "src/stemkit/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["src/stemkit/tables.py", "tables.py"], "kind": "file", "value": "src/stemkit/tables.py"}, {"aliases": [...
[{"answers": ["src/stemkit/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-s1-01290
indexflow
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_footnote_all rejects a range where lo equals hi, which is a valid single point range. It lives in src/indexflow/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 t...
79,691
3,072
114
123
false
synth:5123219254291220485@1.42063
synth
[{"aliases": ["src/indexflow/report.py", "report.py"], "kind": "file", "value": "src/indexflow/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["src/indexflow/format.py", "format.py"], "kind": "file", "value": "src/indexflow/format.py"},...
[{"answers": ["src/indexflow/report.py", "report.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's mo...
ctx_compress-s1-01290-twin
indexflow
ctx_compress-s1-01290
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_footnote_all rejects a range where lo equals hi, which is a valid single point range. It lives in src/indexflow/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 t...
38,035
1,901
62
123
false
synth:5123219254291220485@1.42063
synth
[{"aliases": ["src/indexflow/report.py", "report.py"], "kind": "file", "value": "src/indexflow/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["src/indexflow/format.py", "format.py"], "kind": "file", "value": "src/indexflow/format.py"},...
[{"answers": ["src/indexflow/report.py", "report.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's mo...
ctx_compress-s1-01291
slotworks
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_cadence leaves a trailing dash when the title ends in punctuation. It lives in src/slotworks/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 slotworks layout. [tool...
87,761
3,072
98
109
false
synth:6161565485023161517@1.61472
synth
[{"aliases": ["src/slotworks/cli.py", "cli.py"], "kind": "file", "value": "src/slotworks/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["data/cadence_5.csv", "cadence_5.csv"], "kind": "file", "value": "data/cadence_5.csv"}, {"aliases": ["src/slotw...
[{"answers": ["src/slotworks/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-s1-01291-twin
slotworks
ctx_compress-s1-01291
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_cadence leaves a trailing dash when the title ends in punctuation. It lives in src/slotworks/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 slotworks layout. [tool...
35,815
1,790
54
109
false
synth:6161565485023161517@1.61472
synth
[{"aliases": ["src/slotworks/cli.py", "cli.py"], "kind": "file", "value": "src/slotworks/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["data/cadence_5.csv", "cadence_5.csv"], "kind": "file", "value": "data/cadence_5.csv"}, {"aliases": ["src/slotw...
[{"answers": ["src/slotworks/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-s1-01292
lotflow
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_shipment_strict raises ShipmentShapeError for bytes input in strict mode instead of decoding it. It lives in src/lotflow/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 ma...
73,140
3,072
90
101
false
synth:209174416122076549@1.5244
synth
[{"aliases": ["src/lotflow/core.py", "core.py"], "kind": "file", "value": "src/lotflow/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["data/warehouse_1.csv", "warehouse_1.csv"], "kind": "file", "value": "data/warehouse_1.csv"}, {"aliases": ["d...
[{"answers": ["src/lotflow/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 rec...
ctx_compress-s1-01292-twin
lotflow
ctx_compress-s1-01292
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_shipment_strict raises ShipmentShapeError for bytes input in strict mode instead of decoding it. It lives in src/lotflow/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 ma...
45,559
2,277
56
101
false
synth:209174416122076549@1.5244
synth
[{"aliases": ["src/lotflow/core.py", "core.py"], "kind": "file", "value": "src/lotflow/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["data/warehouse_1.csv", "warehouse_1.csv"], "kind": "file", "value": "data/warehouse_1.csv"}, {"aliases": ["d...
[{"answers": ["src/lotflow/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 rec...
ctx_compress-s1-01293
bearingflow
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_fix raises ValueError when a value itself contains an equals sign. It lives in src/bearingflow/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 bearingflow lay...
55,592
2,779
89
108
false
synth:2779765666919025779@1.22885
synth
[{"aliases": ["src/bearingflow/store.py", "store.py"], "kind": "file", "value": "src/bearingflow/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/bearingflow/schema.py", "schema.py"], "kind": "file", "value": "src/bearingflow/schema.py"...
[{"answers": ["src/bearingflow/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 mo...
ctx_compress-s1-01293-twin
bearingflow
ctx_compress-s1-01293
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_fix raises ValueError when a value itself contains an equals sign. It lives in src/bearingflow/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 bearingflow lay...
29,670
1,483
49
108
false
synth:2779765666919025779@1.22885
synth
[{"aliases": ["src/bearingflow/store.py", "store.py"], "kind": "file", "value": "src/bearingflow/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/bearingflow/schema.py", "schema.py"], "kind": "file", "value": "src/bearingflow/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": ["src/bearingflow/store.py", "store.py"...
ctx_compress-s1-01294
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] extend_principal_batch ignores the cap, so late attempts wait for minutes. It lives in src/scopeledger/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 scopeledger layout. [...
24,900
1,245
51
60
false
synth:494665885476823136@1
synth
[{"aliases": ["src/scopeledger/io.py", "io.py"], "kind": "file", "value": "src/scopeledger/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["data/principal_2.csv", "principal_2.csv"], "kind": "file", "value": "data/principal_2.csv"}, {"aliases": ["src/s...
[{"answers": ["src/scopeledger/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 rec...
ctx_compress-s1-01294-twin
scopeledger
ctx_compress-s1-01294
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_principal_batch ignores the cap, so late attempts wait for minutes. It lives in src/scopeledger/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 scopeledger layout. [...
10,985
1,000
29
60
false
synth:494665885476823136@1
synth
[{"aliases": ["src/scopeledger/io.py", "io.py"], "kind": "file", "value": "src/scopeledger/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["data/principal_2.csv", "principal_2.csv"], "kind": "file", "value": "data/principal_2.csv"}, {"aliases": ["src/s...
[{"answers": ["src/scopeledger/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 rec...
ctx_compress-s1-01295
skukit
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 keeps leading spaces in keys so lookups after a semicolon fail. It lives in src/skukit/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 skukit layout. [tool...
83,886
3,072
89
106
false
synth:5478937510205892889@1.71923
synth
[{"aliases": ["src/skukit/report.py", "report.py"], "kind": "file", "value": "src/skukit/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["data/warehouse_0.csv", "warehouse_0.csv"], "kind": "file", "value": "data/warehouse_0.csv"}, {"ali...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["io.py", "src/skukit/io.py"], "kind": ...
ctx_compress-s1-01295-twin
skukit
ctx_compress-s1-01295
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 keeps leading spaces in keys so lookups after a semicolon fail. It lives in src/skukit/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 skukit layout. [tool...
53,899
2,694
55
106
false
synth:5478937510205892889@1.71923
synth
[{"aliases": ["src/skukit/report.py", "report.py"], "kind": "file", "value": "src/skukit/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["data/warehouse_0.csv", "warehouse_0.csv"], "kind": "file", "value": "data/warehouse_0.csv"}, {"ali...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["data/stock_13.csv", "stock_13.csv"], ...
ctx_compress-s1-01296
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_page double counts carried amounts when the period closes on a weekend. It lives in src/slugflow/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 slugfl...
52,177
2,608
98
107
false
synth:4930486866443427598@1.2983
synth
[{"aliases": ["src/slugflow/store.py", "store.py"], "kind": "file", "value": "src/slugflow/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/slugflow/format.py", "format.py"], "kind": "file", "value": "src/slugflow/format.py"}, {"aliases...
[{"answers": ["src/slugflow/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 ...
ctx_compress-s1-01296-twin
slugflow
ctx_compress-s1-01296
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_page double counts carried amounts when the period closes on a weekend. It lives in src/slugflow/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 slugfl...
35,453
1,772
64
107
false
synth:4930486866443427598@1.2983
synth
[{"aliases": ["src/slugflow/store.py", "store.py"], "kind": "file", "value": "src/slugflow/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/slugflow/format.py", "format.py"], "kind": "file", "value": "src/slugflow/format.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": ["cue_0.csv", "data/cue_0.csv"], "kind"...
ctx_compress-s1-01297
clauseworks
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_clause raises FootnoteShapeError for bytes input in strict mode instead of decoding it. It lives in src/clauseworks/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 map...
87,280
3,072
87
110
false
synth:5987303246612764405@2.3227
synth
[{"aliases": ["src/clauseworks/format.py", "format.py"], "kind": "file", "value": "src/clauseworks/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["src/clauseworks/tables.py", "tables.py"], "kind": "file", "value": "src/clauseworks/tabl...
[{"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-s1-01297-twin
clauseworks
ctx_compress-s1-01297
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_clause raises FootnoteShapeError for bytes input in strict mode instead of decoding it. It lives in src/clauseworks/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 map...
51,468
2,573
53
110
false
synth:5987303246612764405@2.3227
synth
[{"aliases": ["src/clauseworks/format.py", "format.py"], "kind": "file", "value": "src/clauseworks/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["src/clauseworks/tables.py", "tables.py"], "kind": "file", "value": "src/clauseworks/tabl...
[{"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/clauseworks/tables.py", "tables.p...
ctx_compress-s1-01298
spandesk
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] merge_gauge misses keys that differ only in case. It lives in src/spandesk/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 spandesk layout. [tool_use kind=search name...
47,140
2,357
82
99
false
synth:4117970113142347751@1.20348
synth
[{"aliases": ["src/spandesk/rules.py", "rules.py"], "kind": "file", "value": "src/spandesk/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/spandesk/store.py", "store.py"], "kind": "file", "value": "src/spandesk/store.py"}, {"aliases": ...
[{"answers": ["src/spandesk/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-s1-01298-twin
spandesk
ctx_compress-s1-01298
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] merge_gauge misses keys that differ only in case. It lives in src/spandesk/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 spandesk layout. [tool_use kind=search name...
31,907
1,595
54
99
false
synth:4117970113142347751@1.20348
synth
[{"aliases": ["src/spandesk/rules.py", "rules.py"], "kind": "file", "value": "src/spandesk/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/spandesk/store.py", "store.py"], "kind": "file", "value": "src/spandesk/store.py"}, {"aliases": ...
[{"answers": ["src/spandesk/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-s1-01299
nonceworks
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] rotate_grant double counts carried amounts when the period closes on a weekend. It lives in src/nonceworks/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 noncework...
65,731
3,072
103
112
false
synth:78763677325975449@1.44372
synth
[{"aliases": ["src/nonceworks/format.py", "format.py"], "kind": "file", "value": "src/nonceworks/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["src/nonceworks/compat.py", "compat.py"], "kind": "file", "value": "src/nonceworks/compat.p...
[{"answers": ["src/nonceworks/format.py", "format.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's m...
ctx_compress-s1-01299-twin
nonceworks
ctx_compress-s1-01299
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] rotate_grant double counts carried amounts when the period closes on a weekend. It lives in src/nonceworks/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 noncework...
36,227
1,811
55
112
false
synth:78763677325975449@1.44372
synth
[{"aliases": ["src/nonceworks/format.py", "format.py"], "kind": "file", "value": "src/nonceworks/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["src/nonceworks/compat.py", "compat.py"], "kind": "file", "value": "src/nonceworks/compat.p...
[{"answers": ["src/nonceworks/format.py", "format.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's m...
ctx_compress-s1-01300
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] allocate_ledger ignores the cap, so late attempts wait for minutes. 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 ...
87,407
3,072
92
109
false
synth:3844561654837012187@2.0769
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/schema.py", "schema.py"], "kind": "file", "value": "src/ledgerdesk/schema.py"}, {"aliases": ["src/l...
[{"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/ledgerdesk/compat.py...
ctx_compress-s1-01300-twin
ledgerdesk
ctx_compress-s1-01300
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] allocate_ledger ignores the cap, so late attempts wait for minutes. 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 ...
47,182
2,359
61
109
false
synth:3844561654837012187@2.0769
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/schema.py", "schema.py"], "kind": "file", "value": "src/ledgerdesk/schema.py"}, {"aliases": ["src/l...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["rules.py", "src/ledgerdesk/rules.py"]...
ctx_compress-s1-01301
bucketdesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] bin_series_all returns one item too few on every page. It lives in src/bucketdesk/rules.py; `python -m pytest tests/test_rules.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the bucketdesk layout. [tool_use kind=se...
38,523
1,926
70
75
false
synth:1781968973352772332@1
synth
[{"aliases": ["src/bucketdesk/rules.py", "rules.py"], "kind": "file", "value": "src/bucketdesk/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/bucketdesk/schema.py", "schema.py"], "kind": "file", "value": "src/bucketdesk/schema.py"}, {...
[{"answers": ["src/bucketdesk/rules.py", "rules.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's mos...
ctx_compress-s1-01301-twin
bucketdesk
ctx_compress-s1-01301
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_series_all returns one item too few on every page. It lives in src/bucketdesk/rules.py; `python -m pytest tests/test_rules.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the bucketdesk layout. [tool_use kind=se...
15,679
1,000
37
75
false
synth:1781968973352772332@1
synth
[{"aliases": ["src/bucketdesk/rules.py", "rules.py"], "kind": "file", "value": "src/bucketdesk/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/bucketdesk/schema.py", "schema.py"], "kind": "file", "value": "src/bucketdesk/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": ["counter_4.csv", "data/counter_4.csv"]...
ctx_compress-s1-01302
tileforge
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_fix produces doubled dashes for consecutive separators. It lives in src/tileforge/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 tileforge layout. [tool_use kind=sea...
53,894
2,694
76
91
false
synth:6764632660579488056@1.00513
synth
[{"aliases": ["src/tileforge/io.py", "io.py"], "kind": "file", "value": "src/tileforge/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["data/tile_2.csv", "tile_2.csv"], "kind": "file", "value": "data/tile_2.csv"}, {"aliases": ["src/tileforge/report.py"...
[{"answers": ["src/tileforge/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 recen...
ctx_compress-s1-01302-twin
tileforge
ctx_compress-s1-01302
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_fix produces doubled dashes for consecutive separators. It lives in src/tileforge/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 tileforge layout. [tool_use kind=sea...
38,335
1,916
52
91
false
synth:6764632660579488056@1.00513
synth
[{"aliases": ["src/tileforge/io.py", "io.py"], "kind": "file", "value": "src/tileforge/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["data/tile_2.csv", "tile_2.csv"], "kind": "file", "value": "data/tile_2.csv"}, {"aliases": ["src/tileforge/report.py"...
[{"answers": ["src/tileforge/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 recen...
ctx_compress-s1-01303
bearingledger
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_window raises ZeroDivisionError on the first value and lags by one afterwards. It lives in src/bearingledger/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 th...
22,477
1,123
47
54
false
synth:4866086984708046437@1
synth
[{"aliases": ["src/bearingledger/cli.py", "cli.py"], "kind": "file", "value": "src/bearingledger/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["src/bearingledger/format.py", "format.py"], "kind": "file", "value": "src/bearingledger/format.py"}, {...
[{"answers": ["src/bearingledger/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-s1-01303-twin
bearingledger
ctx_compress-s1-01303
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_window raises ZeroDivisionError on the first value and lags by one afterwards. It lives in src/bearingledger/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 th...
7,763
1,000
25
54
false
synth:4866086984708046437@1
synth
[{"aliases": ["src/bearingledger/cli.py", "cli.py"], "kind": "file", "value": "src/bearingledger/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["src/bearingledger/format.py", "format.py"], "kind": "file", "value": "src/bearingledger/format.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": ["test_report.py", "tests/test_report.p...
ctx_compress-s1-01304
bookingkit
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] overlap_shift skips the first page; page 1 returns what should be page 2. It lives in src/bookingkit/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 bookingkit layout. ...
32,116
1,605
61
70
false
synth:6494842787981942368@1
synth
[{"aliases": ["src/bookingkit/core.py", "core.py"], "kind": "file", "value": "src/bookingkit/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/bookingkit/schema.py", "schema.py"], "kind": "file", "value": "src/bookingkit/schema.py"}, {"alias...
[{"answers": ["src/bookingkit/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-s1-01304-twin
bookingkit
ctx_compress-s1-01304
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] overlap_shift skips the first page; page 1 returns what should be page 2. It lives in src/bookingkit/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 bookingkit layout. ...
19,541
1,000
39
70
false
synth:6494842787981942368@1
synth
[{"aliases": ["src/bookingkit/core.py", "core.py"], "kind": "file", "value": "src/bookingkit/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/bookingkit/schema.py", "schema.py"], "kind": "file", "value": "src/bookingkit/schema.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": ["test_io.py", "tests/test_io.py"], "ki...
ctx_compress-s1-01305
counterledger
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_gauge ignores the cap, so late attempts wait for minutes. It lives in src/counterledger/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 counterledger layout. [too...
106,935
3,072
110
139
false
synth:8074333045714664@1.8661
synth
[{"aliases": ["src/counterledger/rules.py", "rules.py"], "kind": "file", "value": "src/counterledger/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["data/trace_7.csv", "trace_7.csv"], "kind": "file", "value": "data/trace_7.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": ["data/probe_4.csv", "probe_4.csv"], "k...
ctx_compress-s1-01305-twin
counterledger
ctx_compress-s1-01305
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_gauge ignores the cap, so late attempts wait for minutes. It lives in src/counterledger/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 counterledger layout. [too...
76,515
3,072
80
139
false
synth:8074333045714664@1.8661
synth
[{"aliases": ["src/counterledger/rules.py", "rules.py"], "kind": "file", "value": "src/counterledger/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["data/trace_7.csv", "trace_7.csv"], "kind": "file", "value": "data/trace_7.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": ["data/gauge_10.csv", "gauge_10.csv"], ...
ctx_compress-s1-01306
indexkit
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_revision leaves a trailing dash when the title ends in punctuation. It lives in src/indexkit/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 indexkit layout....
56,933
2,846
105
116
false
synth:2797295971119267769@1.24585
synth
[{"aliases": ["src/indexkit/compat.py", "compat.py"], "kind": "file", "value": "src/indexkit/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["data/index_0.csv", "index_0.csv"], "kind": "file", "value": "data/index_0.csv"}, {"aliases": [...
[{"answers": ["src/indexkit/compat.py", "compat.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's mos...