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-s1-00503-twin
bearingkit
ctx_compress-s1-00503
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] simplify_bearing raises ZeroDivisionError on the first value and lags by one afterwards. It lives in src/bearingkit/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 bear...
56,798
2,839
70
149
false
synth:6083764731288019384@2.05523
synth
[{"aliases": ["src/bearingkit/core.py", "core.py"], "kind": "file", "value": "src/bearingkit/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["data/fix_4.csv", "fix_4.csv"], "kind": "file", "value": "data/fix_4.csv"}, {"aliases": ["data/tile_5.c...
[{"answers": ["src/bearingkit/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-s1-00504
zoneworks
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_leg_once averages one value too many once the window is full. It lives in src/zoneworks/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 zoneworks layout. [too...
56,145
2,807
118
127
false
synth:3718137479833868700@1.37985
synth
[{"aliases": ["src/zoneworks/store.py", "store.py"], "kind": "file", "value": "src/zoneworks/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/zoneworks/limits.py", "limits.py"], "kind": "file", "value": "src/zoneworks/limits.py"}, {"ali...
[{"answers": ["src/zoneworks/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-s1-00504-twin
zoneworks
ctx_compress-s1-00504
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_leg_once averages one value too many once the window is full. It lives in src/zoneworks/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 zoneworks layout. [too...
38,713
1,935
76
127
false
synth:3718137479833868700@1.37985
synth
[{"aliases": ["src/zoneworks/store.py", "store.py"], "kind": "file", "value": "src/zoneworks/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/zoneworks/limits.py", "limits.py"], "kind": "file", "value": "src/zoneworks/limits.py"}, {"ali...
[{"answers": ["src/zoneworks/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-s1-00505
clausedesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] paginate_margin misses keys that differ only in case. It lives in src/clausedesk/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 clausedesk layout. [tool_use kind=s...
28,911
1,445
48
61
false
synth:2768325612675027156@1
synth
[{"aliases": ["src/clausedesk/limits.py", "limits.py"], "kind": "file", "value": "src/clausedesk/limits.py"}, {"aliases": ["tests/test_limits.py", "test_limits.py"], "kind": "file", "value": "tests/test_limits.py"}, {"aliases": ["src/clausedesk/core.py", "core.py"], "kind": "file", "value": "src/clausedesk/core.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/clausedesk/schema.py...
ctx_compress-s1-00505-twin
clausedesk
ctx_compress-s1-00505
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] paginate_margin misses keys that differ only in case. It lives in src/clausedesk/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 clausedesk layout. [tool_use kind=s...
16,267
1,000
29
61
false
synth:2768325612675027156@1
synth
[{"aliases": ["src/clausedesk/limits.py", "limits.py"], "kind": "file", "value": "src/clausedesk/limits.py"}, {"aliases": ["tests/test_limits.py", "test_limits.py"], "kind": "file", "value": "tests/test_limits.py"}, {"aliases": ["src/clausedesk/core.py", "core.py"], "kind": "file", "value": "src/clausedesk/core.py"}, {...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["io.py", "src/clausedesk/io.py"], "kin...
ctx_compress-s1-00506
sectiondesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] paginate_index averages one value too many once the window is full. It lives in src/sectiondesk/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 sectiondesk layout. ...
48,235
2,411
95
114
false
synth:2758252689262643840@1.44897
synth
[{"aliases": ["src/sectiondesk/limits.py", "limits.py"], "kind": "file", "value": "src/sectiondesk/limits.py"}, {"aliases": ["tests/test_limits.py", "test_limits.py"], "kind": "file", "value": "tests/test_limits.py"}, {"aliases": ["src/sectiondesk/tables.py", "tables.py"], "kind": "file", "value": "src/sectiondesk/tabl...
[{"answers": ["src/sectiondesk/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-s1-00506-twin
sectiondesk
ctx_compress-s1-00506
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] paginate_index averages one value too many once the window is full. It lives in src/sectiondesk/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 sectiondesk layout. ...
35,887
1,794
57
114
false
synth:2758252689262643840@1.44897
synth
[{"aliases": ["src/sectiondesk/limits.py", "limits.py"], "kind": "file", "value": "src/sectiondesk/limits.py"}, {"aliases": ["tests/test_limits.py", "test_limits.py"], "kind": "file", "value": "tests/test_limits.py"}, {"aliases": ["src/sectiondesk/tables.py", "tables.py"], "kind": "file", "value": "src/sectiondesk/tabl...
[{"answers": ["src/sectiondesk/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-s1-00507
sampleforge
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_counter averages one value too many once the window is full. It lives in src/sampleforge/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 sampleforge layout. [too...
70,862
3,072
106
115
false
synth:58748402320060031@1.21313
synth
[{"aliases": ["src/sampleforge/core.py", "core.py"], "kind": "file", "value": "src/sampleforge/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["data/trace_4.csv", "trace_4.csv"], "kind": "file", "value": "data/trace_4.csv"}, {"aliases": ["data/...
[{"answers": ["src/sampleforge/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-s1-00507-twin
sampleforge
ctx_compress-s1-00507
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_counter averages one value too many once the window is full. It lives in src/sampleforge/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 sampleforge layout. [too...
42,453
2,122
64
115
false
synth:58748402320060031@1.21313
synth
[{"aliases": ["src/sampleforge/core.py", "core.py"], "kind": "file", "value": "src/sampleforge/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["data/trace_4.csv", "trace_4.csv"], "kind": "file", "value": "data/trace_4.csv"}, {"aliases": ["data/...
[{"answers": ["src/sampleforge/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-s1-00508
clauseflow
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] renumber_index raises RevisionBoundError for bytes input in strict mode instead of decoding it. It lives in src/clauseflow/compat.py; `python -m pytest tests/test_compat.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mappi...
77,161
3,072
93
118
false
synth:5288946693004382051@1.4194
synth
[{"aliases": ["src/clauseflow/compat.py", "compat.py"], "kind": "file", "value": "src/clauseflow/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["data/revision_4.csv", "revision_4.csv"], "kind": "file", "value": "data/revision_4.csv"}, ...
[{"answers": ["src/clauseflow/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-s1-00508-twin
clauseflow
ctx_compress-s1-00508
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] renumber_index raises RevisionBoundError for bytes input in strict mode instead of decoding it. It lives in src/clauseflow/compat.py; `python -m pytest tests/test_compat.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mappi...
45,536
2,276
65
118
false
synth:5288946693004382051@1.4194
synth
[{"aliases": ["src/clauseflow/compat.py", "compat.py"], "kind": "file", "value": "src/clauseflow/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["data/revision_4.csv", "revision_4.csv"], "kind": "file", "value": "data/revision_4.csv"}, ...
[{"answers": ["src/clauseflow/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-s1-00509
lexemeforge
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] normalise_cue_batch ignores the cap, so late attempts wait for minutes. It lives in src/lexemeforge/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 lexemeforge layout. [too...
76,105
3,072
95
102
false
synth:5054943929765495220@1.16172
synth
[{"aliases": ["src/lexemeforge/io.py", "io.py"], "kind": "file", "value": "src/lexemeforge/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["data/caption_2.csv", "caption_2.csv"], "kind": "file", "value": "data/caption_2.csv"}, {"aliases": ["src/lexemef...
[{"answers": ["src/lexemeforge/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-s1-00509-twin
lexemeforge
ctx_compress-s1-00509
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] normalise_cue_batch ignores the cap, so late attempts wait for minutes. It lives in src/lexemeforge/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 lexemeforge layout. [too...
47,024
2,351
55
102
false
synth:5054943929765495220@1.16172
synth
[{"aliases": ["src/lexemeforge/io.py", "io.py"], "kind": "file", "value": "src/lexemeforge/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["data/caption_2.csv", "caption_2.csv"], "kind": "file", "value": "data/caption_2.csv"}, {"aliases": ["src/lexemef...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["cli.py", "src/lexemeforge/cli.py"], "...
ctx_compress-s1-00510
invoiceforge
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] allocate_tariff double counts carried amounts when the period closes on a weekend. It lives in src/invoiceforge/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 invoic...
83,908
3,072
116
139
false
synth:648507095749682784@1.64795
synth
[{"aliases": ["src/invoiceforge/store.py", "store.py"], "kind": "file", "value": "src/invoiceforge/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/invoiceforge/rules.py", "rules.py"], "kind": "file", "value": "src/invoiceforge/rules.py...
[{"answers": ["src/invoiceforge/store.py", "store.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's m...
ctx_compress-s1-00510-twin
invoiceforge
ctx_compress-s1-00510
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] allocate_tariff double counts carried amounts when the period closes on a weekend. It lives in src/invoiceforge/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 invoic...
45,651
2,282
70
139
false
synth:648507095749682784@1.64795
synth
[{"aliases": ["src/invoiceforge/store.py", "store.py"], "kind": "file", "value": "src/invoiceforge/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/invoiceforge/rules.py", "rules.py"], "kind": "file", "value": "src/invoiceforge/rules.py...
[{"answers": ["src/invoiceforge/store.py", "store.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's m...
ctx_compress-s1-00511
gaugekit
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_trace produces doubled dashes for consecutive separators. It lives in src/gaugekit/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 gaugekit layout. [tool_use kind...
82,961
3,072
116
125
false
synth:3711879830463991967@2.0258
synth
[{"aliases": ["src/gaugekit/core.py", "core.py"], "kind": "file", "value": "src/gaugekit/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/gaugekit/report.py", "report.py"], "kind": "file", "value": "src/gaugekit/report.py"}, {"aliases": ["s...
[{"answers": ["src/gaugekit/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 re...
ctx_compress-s1-00511-twin
gaugekit
ctx_compress-s1-00511
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_trace produces doubled dashes for consecutive separators. It lives in src/gaugekit/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 gaugekit layout. [tool_use kind...
42,159
2,107
72
125
false
synth:3711879830463991967@2.0258
synth
[{"aliases": ["src/gaugekit/core.py", "core.py"], "kind": "file", "value": "src/gaugekit/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/gaugekit/report.py", "report.py"], "kind": "file", "value": "src/gaugekit/report.py"}, {"aliases": ["s...
[{"answers": ["src/gaugekit/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 re...
ctx_compress-s1-00512
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] rotate_scope misses keys that differ only in case. It lives in src/principalforge/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 principalforge layout. [tool_use k...
70,835
3,072
102
121
false
synth:3994842804818086417@1.37815
synth
[{"aliases": ["src/principalforge/compat.py", "compat.py"], "kind": "file", "value": "src/principalforge/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["data/token_5.csv", "token_5.csv"], "kind": "file", "value": "data/token_5.csv"}, {...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["data/nonce_2.csv", "nonce_2.csv"], "k...
ctx_compress-s1-00512-twin
principalforge
ctx_compress-s1-00512
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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 misses keys that differ only in case. It lives in src/principalforge/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 principalforge layout. [tool_use k...
49,752
2,487
68
121
false
synth:3994842804818086417@1.37815
synth
[{"aliases": ["src/principalforge/compat.py", "compat.py"], "kind": "file", "value": "src/principalforge/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["data/token_5.csv", "token_5.csv"], "kind": "file", "value": "data/token_5.csv"}, {...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["test_limits.py", "tests/test_limits.p...
ctx_compress-s1-00513
zoneflow
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] interpolate_waypoint returns the upper bound for every value instead of clamping into the range. It lives in src/zoneflow/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 zo...
38,235
1,911
55
62
false
synth:5463812280115544048@1
synth
[{"aliases": ["src/zoneflow/io.py", "io.py"], "kind": "file", "value": "src/zoneflow/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["src/zoneflow/schema.py", "schema.py"], "kind": "file", "value": "src/zoneflow/schema.py"}, {"aliases": ["data/tile_2.c...
[{"answers": ["src/zoneflow/io.py", "io.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's most recent...
ctx_compress-s1-00513-twin
zoneflow
ctx_compress-s1-00513
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] interpolate_waypoint returns the upper bound for every value instead of clamping into the range. It lives in src/zoneflow/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 zo...
20,261
1,013
33
62
false
synth:5463812280115544048@1
synth
[{"aliases": ["src/zoneflow/io.py", "io.py"], "kind": "file", "value": "src/zoneflow/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["src/zoneflow/schema.py", "schema.py"], "kind": "file", "value": "src/zoneflow/schema.py"}, {"aliases": ["data/tile_2.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": ["src/zoneflow/tables.py", "tables.py"]...
ctx_compress-s1-00514
waypointflow
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_corridor_all keeps leading spaces in keys so lookups after a semicolon fail. It lives in src/waypointflow/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 wa...
52,327
2,616
89
104
false
synth:8885353620372420267@1.33888
synth
[{"aliases": ["src/waypointflow/format.py", "format.py"], "kind": "file", "value": "src/waypointflow/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["docs/changelog.md", "changelog.md"], "kind": "file", "value": "docs/CHANGELOG.md"}, {"...
[{"answers": ["src/waypointflow/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-s1-00514-twin
waypointflow
ctx_compress-s1-00514
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_corridor_all keeps leading spaces in keys so lookups after a semicolon fail. It lives in src/waypointflow/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 wa...
38,326
1,916
61
104
false
synth:8885353620372420267@1.33888
synth
[{"aliases": ["src/waypointflow/format.py", "format.py"], "kind": "file", "value": "src/waypointflow/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["docs/changelog.md", "changelog.md"], "kind": "file", "value": "docs/CHANGELOG.md"}, {"...
[{"answers": ["src/waypointflow/format.py", "format.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's...
ctx_compress-s1-00515
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] revoke_principal grows linearly instead of doubling. It lives in src/nonceworks/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 nonceworks layout. [tool_use kind=search...
22,436
1,121
52
65
false
synth:5459314674923520807@1
synth
[{"aliases": ["src/nonceworks/core.py", "core.py"], "kind": "file", "value": "src/nonceworks/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["data/policy_6.csv", "policy_6.csv"], "kind": "file", "value": "data/policy_6.csv"}, {"aliases": ["src/...
[{"answers": ["src/nonceworks/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-s1-00515-twin
nonceworks
ctx_compress-s1-00515
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] revoke_principal grows linearly instead of doubling. It lives in src/nonceworks/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 nonceworks layout. [tool_use kind=search...
11,414
1,000
34
65
false
synth:5459314674923520807@1
synth
[{"aliases": ["src/nonceworks/core.py", "core.py"], "kind": "file", "value": "src/nonceworks/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["data/policy_6.csv", "policy_6.csv"], "kind": "file", "value": "data/policy_6.csv"}, {"aliases": ["src/...
[{"answers": ["src/nonceworks/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-s1-00516
legworks
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] interpolate_waypoint ignores the cap, so late attempts wait for minutes. It lives in src/legworks/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 legworks layout. [tool_use...
41,943
2,097
81
94
false
synth:4476617818580856971@1
synth
[{"aliases": ["src/legworks/io.py", "io.py"], "kind": "file", "value": "src/legworks/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["src/legworks/tables.py", "tables.py"], "kind": "file", "value": "src/legworks/tables.py"}, {"aliases": ["src/legworks/...
[{"answers": ["src/legworks/io.py", "io.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's most recent...
ctx_compress-s1-00516-twin
legworks
ctx_compress-s1-00516
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] interpolate_waypoint ignores the cap, so late attempts wait for minutes. It lives in src/legworks/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 legworks layout. [tool_use...
17,764
1,000
45
94
false
synth:4476617818580856971@1
synth
[{"aliases": ["src/legworks/io.py", "io.py"], "kind": "file", "value": "src/legworks/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["src/legworks/tables.py", "tables.py"], "kind": "file", "value": "src/legworks/tables.py"}, {"aliases": ["src/legworks/...
[{"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/legworks/compat.py"]...
ctx_compress-s1-00517
cuedesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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 raises ValueError when a value itself contains an equals sign. It lives in src/cuedesk/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 cuedesk layout. [tool_...
78,305
3,072
102
129
false
synth:1204288881922569158@2.22325
synth
[{"aliases": ["src/cuedesk/io.py", "io.py"], "kind": "file", "value": "src/cuedesk/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["src/cuedesk/schema.py", "schema.py"], "kind": "file", "value": "src/cuedesk/schema.py"}, {"aliases": ["docs/changelog.md...
[{"answers": ["src/cuedesk/io.py", "io.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent ...
ctx_compress-s1-00517-twin
cuedesk
ctx_compress-s1-00517
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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 raises ValueError when a value itself contains an equals sign. It lives in src/cuedesk/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 cuedesk layout. [tool_...
62,806
3,072
76
129
false
synth:1204288881922569158@2.22325
synth
[{"aliases": ["src/cuedesk/io.py", "io.py"], "kind": "file", "value": "src/cuedesk/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["src/cuedesk/schema.py", "schema.py"], "kind": "file", "value": "src/cuedesk/schema.py"}, {"aliases": ["docs/changelog.md...
[{"answers": ["src/cuedesk/io.py", "io.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's most recent ...
ctx_compress-s1-00518
tracedesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_gauge produces doubled dashes for consecutive separators. It lives in src/tracedesk/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 tracedesk layout. [tool_use kind=...
106,531
3,072
130
149
false
synth:1790292878512053376@2.29867
synth
[{"aliases": ["src/tracedesk/cli.py", "cli.py"], "kind": "file", "value": "src/tracedesk/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["src/tracedesk/schema.py", "schema.py"], "kind": "file", "value": "src/tracedesk/schema.py"}, {"aliases": ["dat...
[{"answers": ["src/tracedesk/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-s1-00518-twin
tracedesk
ctx_compress-s1-00518
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_gauge produces doubled dashes for consecutive separators. It lives in src/tracedesk/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 tracedesk layout. [tool_use kind=...
70,770
3,072
84
149
false
synth:1790292878512053376@2.29867
synth
[{"aliases": ["src/tracedesk/cli.py", "cli.py"], "kind": "file", "value": "src/tracedesk/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["src/tracedesk/schema.py", "schema.py"], "kind": "file", "value": "src/tracedesk/schema.py"}, {"aliases": ["dat...
[{"answers": ["src/tracedesk/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-s1-00519
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] pick_stock skips the first page; page 1 returns what should be page 2. It lives in src/lotledger/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 lotledger layout. [to...
33,732
1,686
72
83
false
synth:1481056741038176974@1
synth
[{"aliases": ["src/lotledger/store.py", "store.py"], "kind": "file", "value": "src/lotledger/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/lotledger/report.py", "report.py"], "kind": "file", "value": "src/lotledger/report.py"}, {"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": ["src/lotledger/store.py", "store.py"],...
ctx_compress-s1-00519-twin
lotledger
ctx_compress-s1-00519
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] pick_stock skips the first page; page 1 returns what should be page 2. It lives in src/lotledger/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 lotledger layout. [to...
13,199
1,000
39
83
false
synth:1481056741038176974@1
synth
[{"aliases": ["src/lotledger/store.py", "store.py"], "kind": "file", "value": "src/lotledger/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/lotledger/report.py", "report.py"], "kind": "file", "value": "src/lotledger/report.py"}, {"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": ["src/lotledger/tables.py", "tables.py"...
ctx_compress-s1-00520
indexdesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] paginate_section_safe misses keys that differ only in case. It lives in src/indexdesk/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 indexdesk layout. [tool_use ki...
32,787
1,639
60
69
false
synth:2081025321283966210@1
synth
[{"aliases": ["src/indexdesk/report.py", "report.py"], "kind": "file", "value": "src/indexdesk/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["src/indexdesk/compat.py", "compat.py"], "kind": "file", "value": "src/indexdesk/compat.py"},...
[{"answers": ["src/indexdesk/report.py", "report.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's mo...
ctx_compress-s1-00520-twin
indexdesk
ctx_compress-s1-00520
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] paginate_section_safe misses keys that differ only in case. It lives in src/indexdesk/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 indexdesk layout. [tool_use ki...
16,761
1,000
39
69
false
synth:2081025321283966210@1
synth
[{"aliases": ["src/indexdesk/report.py", "report.py"], "kind": "file", "value": "src/indexdesk/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["src/indexdesk/compat.py", "compat.py"], "kind": "file", "value": "src/indexdesk/compat.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": ["report.py", "src/indexdesk/report.py"...
ctx_compress-s1-00521
fixledger
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] simplify_tile ignores the cap, so late attempts wait for minutes. It lives in src/fixledger/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 fixledger layout. [tool_use ...
54,093
2,704
78
99
false
synth:1818170561555329263@1.82935
synth
[{"aliases": ["src/fixledger/core.py", "core.py"], "kind": "file", "value": "src/fixledger/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/fixledger/tables.py", "tables.py"], "kind": "file", "value": "src/fixledger/tables.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": ["changelog.md", "docs/changelog.md"], ...
ctx_compress-s1-00521-twin
fixledger
ctx_compress-s1-00521
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] simplify_tile ignores the cap, so late attempts wait for minutes. It lives in src/fixledger/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 fixledger layout. [tool_use ...
35,089
1,754
55
99
false
synth:1818170561555329263@1.82935
synth
[{"aliases": ["src/fixledger/core.py", "core.py"], "kind": "file", "value": "src/fixledger/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/fixledger/tables.py", "tables.py"], "kind": "file", "value": "src/fixledger/tables.py"}, {"aliases":...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["data/tile_4.csv", "tile_4.csv"], "kin...
ctx_compress-s1-00522
seriesforge
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] rate_span_once drops override values of 0, False and empty string. It lives in src/seriesforge/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 seriesforge layout. [...
75,334
3,072
105
116
false
synth:6232041706788453419@2.0902
synth
[{"aliases": ["src/seriesforge/report.py", "report.py"], "kind": "file", "value": "src/seriesforge/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["data/trace_7.csv", "trace_7.csv"], "kind": "file", "value": "data/trace_7.csv"}, {"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": ["report.py", "src/seriesforge/report.p...
ctx_compress-s1-00522-twin
seriesforge
ctx_compress-s1-00522
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] rate_span_once drops override values of 0, False and empty string. It lives in src/seriesforge/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 seriesforge layout. [...
35,922
1,796
53
116
false
synth:6232041706788453419@2.0902
synth
[{"aliases": ["src/seriesforge/report.py", "report.py"], "kind": "file", "value": "src/seriesforge/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["data/trace_7.csv", "trace_7.csv"], "kind": "file", "value": "data/trace_7.csv"}, {"alias...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["data/series_1.csv", "series_1.csv"], ...
ctx_compress-s1-00523
seriesledger
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_counter produces doubled dashes for consecutive separators. It lives in src/seriesledger/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 seriesledger layout. [tool_...
27,958
1,397
69
84
false
synth:6231061549265960754@1
synth
[{"aliases": ["src/seriesledger/cli.py", "cli.py"], "kind": "file", "value": "src/seriesledger/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["src/seriesledger/format.py", "format.py"], "kind": "file", "value": "src/seriesledger/format.py"}, {"ali...
[{"answers": ["src/seriesledger/cli.py", "cli.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most ...
ctx_compress-s1-00523-twin
seriesledger
ctx_compress-s1-00523
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_counter produces doubled dashes for consecutive separators. It lives in src/seriesledger/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 seriesledger layout. [tool_...
13,301
1,000
43
84
false
synth:6231061549265960754@1
synth
[{"aliases": ["src/seriesledger/cli.py", "cli.py"], "kind": "file", "value": "src/seriesledger/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["src/seriesledger/format.py", "format.py"], "kind": "file", "value": "src/seriesledger/format.py"}, {"ali...
[{"answers": ["src/seriesledger/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-s1-00524
indexdesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] collate_clause_batch lets defaults overwrite explicit overrides. It lives in src/indexdesk/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 indexdesk layout. [tool_use...
67,349
3,072
91
116
false
synth:3994676647608741803@1.67237
synth
[{"aliases": ["src/indexdesk/rules.py", "rules.py"], "kind": "file", "value": "src/indexdesk/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/indexdesk/limits.py", "limits.py"], "kind": "file", "value": "src/indexdesk/limits.py"}, {"ali...
[{"answers": ["src/indexdesk/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-s1-00524-twin
indexdesk
ctx_compress-s1-00524
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_clause_batch lets defaults overwrite explicit overrides. It lives in src/indexdesk/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 indexdesk layout. [tool_use...
48,834
2,441
69
116
false
synth:3994676647608741803@1.67237
synth
[{"aliases": ["src/indexdesk/rules.py", "rules.py"], "kind": "file", "value": "src/indexdesk/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/indexdesk/limits.py", "limits.py"], "kind": "file", "value": "src/indexdesk/limits.py"}, {"ali...
[{"answers": ["src/indexdesk/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-s1-00525
zonedesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_leg misses keys that differ only in case. It lives in src/zonedesk/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 zonedesk layout. [tool_use kind=search name=B...
54,207
2,710
88
109
false
synth:3270246248313845069@1.45355
synth
[{"aliases": ["src/zonedesk/audit.py", "audit.py"], "kind": "file", "value": "src/zonedesk/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["src/zonedesk/io.py", "io.py"], "kind": "file", "value": "src/zonedesk/io.py"}, {"aliases": ["data/fi...
[{"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/tile_1.csv", "tile_1.csv"], "kin...
ctx_compress-s1-00525-twin
zonedesk
ctx_compress-s1-00525
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_leg misses keys that differ only in case. It lives in src/zonedesk/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 zonedesk layout. [tool_use kind=search name=B...
42,343
2,117
64
109
false
synth:3270246248313845069@1.45355
synth
[{"aliases": ["src/zonedesk/audit.py", "audit.py"], "kind": "file", "value": "src/zonedesk/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["src/zonedesk/io.py", "io.py"], "kind": "file", "value": "src/zonedesk/io.py"}, {"aliases": ["data/fi...
[{"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-s1-00526
lexemekit
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_stem drops override values of 0, False and empty string. It lives in src/lexemekit/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 lexemekit layout. [tool_...
64,901
3,072
88
101
false
synth:4281179600027708132@1.16575
synth
[{"aliases": ["src/lexemekit/format.py", "format.py"], "kind": "file", "value": "src/lexemekit/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["src/lexemekit/limits.py", "limits.py"], "kind": "file", "value": "src/lexemekit/limits.py"},...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["schema.py", "src/lexemekit/schema.py"...
ctx_compress-s1-00526-twin
lexemekit
ctx_compress-s1-00526
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_stem drops override values of 0, False and empty string. It lives in src/lexemekit/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 lexemekit layout. [tool_...
36,870
1,843
56
101
false
synth:4281179600027708132@1.16575
synth
[{"aliases": ["src/lexemekit/format.py", "format.py"], "kind": "file", "value": "src/lexemekit/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["src/lexemekit/limits.py", "limits.py"], "kind": "file", "value": "src/lexemekit/limits.py"},...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["data/slug_4.csv", "slug_4.csv"], "kin...
ctx_compress-s1-00527
shiftkit
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_booking_window raises BookingStateError for bytes input in strict mode instead of decoding it. It lives in src/shiftkit/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 mappin...
54,841
2,742
83
90
false
synth:560437486514167456@1
synth
[{"aliases": ["src/shiftkit/core.py", "core.py"], "kind": "file", "value": "src/shiftkit/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/shiftkit/store.py", "store.py"], "kind": "file", "value": "src/shiftkit/store.py"}, {"aliases": ["src/...
[{"answers": ["src/shiftkit/core.py", "core.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's most re...
ctx_compress-s1-00527-twin
shiftkit
ctx_compress-s1-00527
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_booking_window raises BookingStateError for bytes input in strict mode instead of decoding it. It lives in src/shiftkit/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 mappin...
30,147
1,507
49
90
false
synth:560437486514167456@1
synth
[{"aliases": ["src/shiftkit/core.py", "core.py"], "kind": "file", "value": "src/shiftkit/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/shiftkit/store.py", "store.py"], "kind": "file", "value": "src/shiftkit/store.py"}, {"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-s1-00528
revisionledger
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] anchor_clause misses keys that differ only in case. It lives in src/revisionledger/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 revisionledger layout. [tool_use ki...
69,796
3,072
101
108
false
synth:3573131436539668609@1.49195
synth
[{"aliases": ["src/revisionledger/store.py", "store.py"], "kind": "file", "value": "src/revisionledger/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/revisionledger/tables.py", "tables.py"], "kind": "file", "value": "src/revisionledge...
[{"answers": ["src/revisionledger/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...
ctx_compress-s1-00528-twin
revisionledger
ctx_compress-s1-00528
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] anchor_clause misses keys that differ only in case. It lives in src/revisionledger/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 revisionledger layout. [tool_use ki...
39,956
1,997
57
108
false
synth:3573131436539668609@1.49195
synth
[{"aliases": ["src/revisionledger/store.py", "store.py"], "kind": "file", "value": "src/revisionledger/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/revisionledger/tables.py", "tables.py"], "kind": "file", "value": "src/revisionledge...
[{"answers": ["src/revisionledger/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...
ctx_compress-s1-00529
sampledesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_trace skips the first page; page 1 returns what should be page 2. It lives in src/sampledesk/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 sampledesk layout. [...
113,812
3,072
136
153
false
synth:6293381070175468876@2.13932
synth
[{"aliases": ["src/sampledesk/store.py", "store.py"], "kind": "file", "value": "src/sampledesk/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/sampledesk/tables.py", "tables.py"], "kind": "file", "value": "src/sampledesk/tables.py"}, {...
[{"answers": ["src/sampledesk/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-s1-00529-twin
sampledesk
ctx_compress-s1-00529
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_trace skips the first page; page 1 returns what should be page 2. It lives in src/sampledesk/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 sampledesk layout. [...
63,314
3,072
82
153
false
synth:6293381070175468876@2.13932
synth
[{"aliases": ["src/sampledesk/store.py", "store.py"], "kind": "file", "value": "src/sampledesk/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/sampledesk/tables.py", "tables.py"], "kind": "file", "value": "src/sampledesk/tables.py"}, {...
[{"answers": ["src/sampledesk/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-s1-00530
indexdesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] renumber_clause_window misses keys that differ only in case. It lives in src/indexdesk/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 indexdesk layout. [tool_use kin...
60,070
3,003
99
106
false
synth:1239298460265159605@1.2217
synth
[{"aliases": ["src/indexdesk/store.py", "store.py"], "kind": "file", "value": "src/indexdesk/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/indexdesk/tables.py", "tables.py"], "kind": "file", "value": "src/indexdesk/tables.py"}, {"ali...
[{"answers": ["src/indexdesk/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-s1-00530-twin
indexdesk
ctx_compress-s1-00530
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_window misses keys that differ only in case. It lives in src/indexdesk/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 indexdesk layout. [tool_use kin...
33,147
1,657
61
106
false
synth:1239298460265159605@1.2217
synth
[{"aliases": ["src/indexdesk/store.py", "store.py"], "kind": "file", "value": "src/indexdesk/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/indexdesk/tables.py", "tables.py"], "kind": "file", "value": "src/indexdesk/tables.py"}, {"ali...
[{"answers": ["src/indexdesk/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-s1-00531
traceledger
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_series ignores the cap, so late attempts wait for minutes. It lives in src/traceledger/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 traceledger layout. [tool...
41,839
2,091
87
100
false
synth:3321370865379847197@1.49765
synth
[{"aliases": ["src/traceledger/rules.py", "rules.py"], "kind": "file", "value": "src/traceledger/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["data/trace_1.csv", "trace_1.csv"], "kind": "file", "value": "data/trace_1.csv"}, {"aliases": [...
[{"answers": ["src/traceledger/rules.py", "rules.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's mo...
ctx_compress-s1-00531-twin
traceledger
ctx_compress-s1-00531
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_series ignores the cap, so late attempts wait for minutes. It lives in src/traceledger/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 traceledger layout. [tool...
27,767
1,388
59
100
false
synth:3321370865379847197@1.49765
synth
[{"aliases": ["src/traceledger/rules.py", "rules.py"], "kind": "file", "value": "src/traceledger/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["data/trace_1.csv", "trace_1.csv"], "kind": "file", "value": "data/trace_1.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": ["test_core.py", "tests/test_core.py"],...
ctx_compress-s1-00532
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] fold_index rejects a range where lo equals hi, which is a valid single point range. It lives in src/indexflow/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 indexflow ...
27,146
1,357
70
83
false
synth:2229966629744976957@1
synth
[{"aliases": ["src/indexflow/core.py", "core.py"], "kind": "file", "value": "src/indexflow/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/indexflow/format.py", "format.py"], "kind": "file", "value": "src/indexflow/format.py"}, {"aliases":...
[{"answers": ["src/indexflow/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-s1-00532-twin
indexflow
ctx_compress-s1-00532
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] fold_index rejects a range where lo equals hi, which is a valid single point range. It lives in src/indexflow/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 indexflow ...
14,968
1,000
40
83
false
synth:2229966629744976957@1
synth
[{"aliases": ["src/indexflow/core.py", "core.py"], "kind": "file", "value": "src/indexflow/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/indexflow/format.py", "format.py"], "kind": "file", "value": "src/indexflow/format.py"}, {"aliases":...
[{"answers": ["src/indexflow/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 r...
ctx_compress-s1-00533
corridorledger
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] interpolate_corridor raises TileShapeError for bytes input in strict mode instead of decoding it. It lives in src/corridorledger/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...
109,553
3,072
119
130
false
synth:141809623987119727@1.90572
synth
[{"aliases": ["src/corridorledger/compat.py", "compat.py"], "kind": "file", "value": "src/corridorledger/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["data/beacon_10.csv", "beacon_10.csv"], "kind": "file", "value": "data/beacon_10.cs...
[{"answers": ["src/corridorledger/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...
ctx_compress-s1-00533-twin
corridorledger
ctx_compress-s1-00533
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] interpolate_corridor raises TileShapeError for bytes input in strict mode instead of decoding it. It lives in src/corridorledger/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...
61,892
3,072
73
130
false
synth:141809623987119727@1.90572
synth
[{"aliases": ["src/corridorledger/compat.py", "compat.py"], "kind": "file", "value": "src/corridorledger/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["data/beacon_10.csv", "beacon_10.csv"], "kind": "file", "value": "data/beacon_10.cs...
[{"answers": ["src/corridorledger/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...
ctx_compress-s1-00534
headingledger
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_margin raises MarginShapeError for bytes input in strict mode instead of decoding it. It lives in src/headingledger/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...
38,788
1,939
67
74
false
synth:1050541656170398946@1
synth
[{"aliases": ["src/headingledger/cli.py", "cli.py"], "kind": "file", "value": "src/headingledger/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["src/headingledger/report.py", "report.py"], "kind": "file", "value": "src/headingledger/report.py"}, {...
[{"answers": ["src/headingledger/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-s1-00534-twin
headingledger
ctx_compress-s1-00534
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_margin raises MarginShapeError for bytes input in strict mode instead of decoding it. It lives in src/headingledger/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...
12,649
1,000
37
74
false
synth:1050541656170398946@1
synth
[{"aliases": ["src/headingledger/cli.py", "cli.py"], "kind": "file", "value": "src/headingledger/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["src/headingledger/report.py", "report.py"], "kind": "file", "value": "src/headingledger/report.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/section_5.csv", "section_5.csv"]...
ctx_compress-s1-00535
rosterforge
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] rotate_cadence_batch double counts carried amounts when the period closes on a weekend. It lives in src/rosterforge/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 ro...
62,289
3,072
86
103
false
synth:7762273616092724222@1.286
synth
[{"aliases": ["src/rosterforge/rules.py", "rules.py"], "kind": "file", "value": "src/rosterforge/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/rosterforge/store.py", "store.py"], "kind": "file", "value": "src/rosterforge/store.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": ["format.py", "src/rosterforge/format.p...
ctx_compress-s1-00535-twin
rosterforge
ctx_compress-s1-00535
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] rotate_cadence_batch double counts carried amounts when the period closes on a weekend. It lives in src/rosterforge/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 ro...
28,661
1,433
51
103
false
synth:7762273616092724222@1.286
synth
[{"aliases": ["src/rosterforge/rules.py", "rules.py"], "kind": "file", "value": "src/rosterforge/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/rosterforge/store.py", "store.py"], "kind": "file", "value": "src/rosterforge/store.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": ["cadence_3.csv", "data/cadence_3.csv"]...
ctx_compress-s1-00536
sectiondesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] paginate_draft_batch drops override values of 0, False and empty string. It lives in src/sectiondesk/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 sectiondesk lay...
62,829
3,072
84
107
false
synth:2432115361589724794@1.99558
synth
[{"aliases": ["src/sectiondesk/format.py", "format.py"], "kind": "file", "value": "src/sectiondesk/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["data/draft_5.csv", "draft_5.csv"], "kind": "file", "value": "data/draft_5.csv"}, {"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": ["schema.py", "src/sectiondesk/schema.p...
ctx_compress-s1-00536-twin
sectiondesk
ctx_compress-s1-00536
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] paginate_draft_batch drops override values of 0, False and empty string. It lives in src/sectiondesk/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 sectiondesk lay...
43,860
2,193
62
107
false
synth:2432115361589724794@1.99558
synth
[{"aliases": ["src/sectiondesk/format.py", "format.py"], "kind": "file", "value": "src/sectiondesk/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["data/draft_5.csv", "draft_5.csv"], "kind": "file", "value": "data/draft_5.csv"}, {"alias...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["data/draft_2.csv", "draft_2.csv"], "k...
ctx_compress-s1-00537
seriesflow
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] downsample_counter returns one item too few on every page. It lives in src/seriesflow/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 seriesflow layout. [tool_use kin...
67,284
3,072
90
113
false
synth:6021155266987045503@1.96853
synth
[{"aliases": ["src/seriesflow/rules.py", "rules.py"], "kind": "file", "value": "src/seriesflow/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/seriesflow/limits.py", "limits.py"], "kind": "file", "value": "src/seriesflow/limits.py"}, {...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["test_core.py", "tests/test_core.py"],...
ctx_compress-s1-00537-twin
seriesflow
ctx_compress-s1-00537
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] downsample_counter returns one item too few on every page. It lives in src/seriesflow/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 seriesflow layout. [tool_use kin...
46,526
2,326
62
113
false
synth:6021155266987045503@1.96853
synth
[{"aliases": ["src/seriesflow/rules.py", "rules.py"], "kind": "file", "value": "src/seriesflow/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/seriesflow/limits.py", "limits.py"], "kind": "file", "value": "src/seriesflow/limits.py"}, {...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["data/gauge_1.csv", "gauge_1.csv"], "k...
ctx_compress-s1-00538
beaconkit
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] project_fix raises ValueError when a value itself contains an equals sign. It lives in src/beaconkit/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 beaconkit layou...
58,859
2,942
89
98
false
synth:6063140229583811859@1
synth
[{"aliases": ["src/beaconkit/report.py", "report.py"], "kind": "file", "value": "src/beaconkit/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["src/beaconkit/rules.py", "rules.py"], "kind": "file", "value": "src/beaconkit/rules.py"}, {"...
[{"answers": ["src/beaconkit/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-s1-00538-twin
beaconkit
ctx_compress-s1-00538
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] project_fix raises ValueError when a value itself contains an equals sign. It lives in src/beaconkit/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 beaconkit layou...
30,605
1,530
51
98
false
synth:6063140229583811859@1
synth
[{"aliases": ["src/beaconkit/report.py", "report.py"], "kind": "file", "value": "src/beaconkit/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["src/beaconkit/rules.py", "rules.py"], "kind": "file", "value": "src/beaconkit/rules.py"}, {"...
[{"answers": ["src/beaconkit/report.py", "report.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's mo...
ctx_compress-s1-00539
cueworks
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] hyphenate_chunk averages one value too many once the window is full. It lives in src/cueworks/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 cueworks layout. [tool_use kin...
86,773
3,072
104
121
false
synth:9219376449161705689@1.87888
synth
[{"aliases": ["src/cueworks/io.py", "io.py"], "kind": "file", "value": "src/cueworks/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["data/cue_2.csv", "cue_2.csv"], "kind": "file", "value": "data/cue_2.csv"}, {"aliases": ["data/chunk_0.csv", "chunk_0.c...
[{"answers": ["src/cueworks/io.py", "io.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent...
ctx_compress-s1-00539-twin
cueworks
ctx_compress-s1-00539
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] hyphenate_chunk averages one value too many once the window is full. It lives in src/cueworks/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 cueworks layout. [tool_use kin...
42,296
2,114
55
121
false
synth:9219376449161705689@1.87888
synth
[{"aliases": ["src/cueworks/io.py", "io.py"], "kind": "file", "value": "src/cueworks/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["data/cue_2.csv", "cue_2.csv"], "kind": "file", "value": "data/cue_2.csv"}, {"aliases": ["data/chunk_0.csv", "chunk_0.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": ["cue_6.csv", "data/cue_6.csv"], "kind"...
ctx_compress-s1-00540
shiftdesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_quota leaves a trailing dash when the title ends in punctuation. It lives in src/shiftdesk/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 shiftdesk layout. [tool_us...
68,712
3,072
86
107
false
synth:2121703209293788570@2.01805
synth
[{"aliases": ["src/shiftdesk/io.py", "io.py"], "kind": "file", "value": "src/shiftdesk/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["src/shiftdesk/schema.py", "schema.py"], "kind": "file", "value": "src/shiftdesk/schema.py"}, {"aliases": ["src/shift...
[{"answers": ["src/shiftdesk/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-s1-00540-twin
shiftdesk
ctx_compress-s1-00540
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_quota leaves a trailing dash when the title ends in punctuation. It lives in src/shiftdesk/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 shiftdesk layout. [tool_us...
29,881
1,494
51
107
false
synth:2121703209293788570@2.01805
synth
[{"aliases": ["src/shiftdesk/io.py", "io.py"], "kind": "file", "value": "src/shiftdesk/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["src/shiftdesk/schema.py", "schema.py"], "kind": "file", "value": "src/shiftdesk/schema.py"}, {"aliases": ["src/shift...
[{"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/slot_2.csv", "slot_2.csv"], "kin...
ctx_compress-s1-00541
probeforge
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_span leaves a trailing dash when the title ends in punctuation. It lives in src/probeforge/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 probeforge layout. [tool...
74,991
3,072
76
97
false
synth:7491019888533248933@1.65175
synth
[{"aliases": ["src/probeforge/core.py", "core.py"], "kind": "file", "value": "src/probeforge/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/probeforge/rules.py", "rules.py"], "kind": "file", "value": "src/probeforge/rules.py"}, {"aliases"...
[{"answers": ["src/probeforge/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-s1-00541-twin
probeforge
ctx_compress-s1-00541
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_span leaves a trailing dash when the title ends in punctuation. It lives in src/probeforge/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 probeforge layout. [tool...
44,957
2,247
56
97
false
synth:7491019888533248933@1.65175
synth
[{"aliases": ["src/probeforge/core.py", "core.py"], "kind": "file", "value": "src/probeforge/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/probeforge/rules.py", "rules.py"], "kind": "file", "value": "src/probeforge/rules.py"}, {"aliases"...
[{"answers": ["src/probeforge/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-s1-00542
calendarkit
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] balance_cadence produces doubled dashes for consecutive separators. It lives in src/calendarkit/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 calendarkit layout. ...
18,774
1,000
41
46
false
synth:3002684176615009903@1
synth
[{"aliases": ["src/calendarkit/compat.py", "compat.py"], "kind": "file", "value": "src/calendarkit/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["data/booking_2.csv", "booking_2.csv"], "kind": "file", "value": "data/booking_2.csv"}, {...
[{"answers": ["src/calendarkit/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-s1-00542-twin
calendarkit
ctx_compress-s1-00542
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] balance_cadence produces doubled dashes for consecutive separators. It lives in src/calendarkit/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 calendarkit layout. ...
10,828
1,000
27
46
false
synth:3002684176615009903@1
synth
[{"aliases": ["src/calendarkit/compat.py", "compat.py"], "kind": "file", "value": "src/calendarkit/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["data/booking_2.csv", "booking_2.csv"], "kind": "file", "value": "data/booking_2.csv"}, {...
[{"answers": ["src/calendarkit/compat.py", "compat.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's ...
ctx_compress-s1-00543
balanceflow
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] settle_ledger raises ZeroDivisionError on the first value and lags by one afterwards. It lives in src/balanceflow/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 ba...
65,843
3,072
96
111
false
synth:272310844458113035@2.2118
synth
[{"aliases": ["src/balanceflow/limits.py", "limits.py"], "kind": "file", "value": "src/balanceflow/limits.py"}, {"aliases": ["tests/test_limits.py", "test_limits.py"], "kind": "file", "value": "tests/test_limits.py"}, {"aliases": ["data/balance_1.csv", "balance_1.csv"], "kind": "file", "value": "data/balance_1.csv"}, {...
[{"answers": ["src/balanceflow/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-s1-00543-twin
balanceflow
ctx_compress-s1-00543
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] settle_ledger raises ZeroDivisionError on the first value and lags by one afterwards. It lives in src/balanceflow/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 ba...
42,983
2,149
61
111
false
synth:272310844458113035@2.2118
synth
[{"aliases": ["src/balanceflow/limits.py", "limits.py"], "kind": "file", "value": "src/balanceflow/limits.py"}, {"aliases": ["tests/test_limits.py", "test_limits.py"], "kind": "file", "value": "tests/test_limits.py"}, {"aliases": ["data/balance_1.csv", "balance_1.csv"], "kind": "file", "value": "data/balance_1.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": ["limits.py", "src/balanceflow/limits.p...
ctx_compress-s1-00544
palletdesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] reserve_warehouse double counts carried amounts when the period closes on a weekend. It lives in src/palletdesk/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 pallet...
32,250
1,612
58
65
false
synth:5320478878291903626@1
synth
[{"aliases": ["src/palletdesk/store.py", "store.py"], "kind": "file", "value": "src/palletdesk/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/palletdesk/audit.py", "audit.py"], "kind": "file", "value": "src/palletdesk/audit.py"}, {"al...
[{"answers": ["src/palletdesk/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-s1-00544-twin
palletdesk
ctx_compress-s1-00544
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_warehouse double counts carried amounts when the period closes on a weekend. It lives in src/palletdesk/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 pallet...
14,581
1,000
34
65
false
synth:5320478878291903626@1
synth
[{"aliases": ["src/palletdesk/store.py", "store.py"], "kind": "file", "value": "src/palletdesk/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/palletdesk/audit.py", "audit.py"], "kind": "file", "value": "src/palletdesk/audit.py"}, {"al...
[{"answers": ["src/palletdesk/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-s1-00545
receiptforge
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_credit drops override values of 0, False and empty string. It lives in src/receiptforge/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 receiptforge layout. [to...
112,598
3,072
126
145
false
synth:1869945334204121985@2.12493
synth
[{"aliases": ["src/receiptforge/cli.py", "cli.py"], "kind": "file", "value": "src/receiptforge/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["docs/changelog.md", "changelog.md"], "kind": "file", "value": "docs/CHANGELOG.md"}, {"aliases": ["data/t...
[{"answers": ["src/receiptforge/cli.py", "cli.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most ...
ctx_compress-s1-00545-twin
receiptforge
ctx_compress-s1-00545
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_credit drops override values of 0, False and empty string. It lives in src/receiptforge/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 receiptforge layout. [to...
62,312
3,072
80
145
false
synth:1869945334204121985@2.12493
synth
[{"aliases": ["src/receiptforge/cli.py", "cli.py"], "kind": "file", "value": "src/receiptforge/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["docs/changelog.md", "changelog.md"], "kind": "file", "value": "docs/CHANGELOG.md"}, {"aliases": ["data/t...
[{"answers": ["src/receiptforge/cli.py", "cli.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most ...
ctx_compress-s1-00546
slotkit
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_window misses keys that differ only in case. It lives in src/slotkit/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 slotkit layout. [tool_use kind=search n...
39,439
1,971
84
91
false
synth:8594762385297656558@1
synth
[{"aliases": ["src/slotkit/report.py", "report.py"], "kind": "file", "value": "src/slotkit/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["src/slotkit/audit.py", "audit.py"], "kind": "file", "value": "src/slotkit/audit.py"}, {"aliases"...
[{"answers": ["src/slotkit/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 most...
ctx_compress-s1-00546-twin
slotkit
ctx_compress-s1-00546
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_window misses keys that differ only in case. It lives in src/slotkit/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 slotkit layout. [tool_use kind=search n...
19,108
1,000
42
91
false
synth:8594762385297656558@1
synth
[{"aliases": ["src/slotkit/report.py", "report.py"], "kind": "file", "value": "src/slotkit/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["src/slotkit/audit.py", "audit.py"], "kind": "file", "value": "src/slotkit/audit.py"}, {"aliases"...
[{"answers": ["src/slotkit/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 most...
ctx_compress-s1-00547
waypointworks
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_corridor returns the upper bound for every value instead of clamping into the range. It lives in src/waypointworks/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 ...
84,557
3,072
106
117
false
synth:6124247575625602226@1.6851
synth
[{"aliases": ["src/waypointworks/core.py", "core.py"], "kind": "file", "value": "src/waypointworks/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["data/tile_0.csv", "tile_0.csv"], "kind": "file", "value": "data/tile_0.csv"}, {"aliases": ["data...
[{"answers": ["src/waypointworks/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 mo...
ctx_compress-s1-00547-twin
waypointworks
ctx_compress-s1-00547
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_corridor returns the upper bound for every value instead of clamping into the range. It lives in src/waypointworks/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 ...
49,873
2,493
64
117
false
synth:6124247575625602226@1.6851
synth
[{"aliases": ["src/waypointworks/core.py", "core.py"], "kind": "file", "value": "src/waypointworks/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["data/tile_0.csv", "tile_0.csv"], "kind": "file", "value": "data/tile_0.csv"}, {"aliases": ["data...
[{"answers": ["src/waypointworks/core.py", "core.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's mo...
ctx_compress-s1-00548
captionforge
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_slug_strict leaves a trailing dash when the title ends in punctuation. It lives in src/captionforge/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 captionforge la...
73,951
3,072
109
134
false
synth:1688613037318061473@1.54482
synth
[{"aliases": ["src/captionforge/cli.py", "cli.py"], "kind": "file", "value": "src/captionforge/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["data/cue_4.csv", "cue_4.csv"], "kind": "file", "value": "data/cue_4.csv"}, {"aliases": ["src/captionforg...
[{"answers": ["src/captionforge/cli.py", "cli.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most ...
ctx_compress-s1-00548-twin
captionforge
ctx_compress-s1-00548
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_slug_strict leaves a trailing dash when the title ends in punctuation. It lives in src/captionforge/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 captionforge la...
47,614
2,380
73
134
false
synth:1688613037318061473@1.54482
synth
[{"aliases": ["src/captionforge/cli.py", "cli.py"], "kind": "file", "value": "src/captionforge/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["data/cue_4.csv", "cue_4.csv"], "kind": "file", "value": "data/cue_4.csv"}, {"aliases": ["src/captionforg...
[{"answers": ["src/captionforge/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-s1-00549
serieskit
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_span raises KeyError for a missing key instead of returning the default. It lives in src/serieskit/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 serieskit layo...
59,371
2,968
97
118
false
synth:5443983925384971606@1.73382
synth
[{"aliases": ["src/serieskit/audit.py", "audit.py"], "kind": "file", "value": "src/serieskit/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["src/serieskit/tables.py", "tables.py"], "kind": "file", "value": "src/serieskit/tables.py"}, {"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": ["test_cli.py", "tests/test_cli.py"], "...
ctx_compress-s1-00549-twin
serieskit
ctx_compress-s1-00549
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_span raises KeyError for a missing key instead of returning the default. It lives in src/serieskit/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 serieskit layo...
27,230
1,361
55
118
false
synth:5443983925384971606@1.73382
synth
[{"aliases": ["src/serieskit/audit.py", "audit.py"], "kind": "file", "value": "src/serieskit/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["src/serieskit/tables.py", "tables.py"], "kind": "file", "value": "src/serieskit/tables.py"}, {"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": ["bucket_2.csv", "data/bucket_2.csv"], ...
ctx_compress-s1-00550
calendarflow
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_window lets defaults overwrite explicit overrides. It lives in src/calendarflow/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 calendarflow layout. [tool_use ki...
36,562
1,828
66
81
false
synth:583569567764500037@1
synth
[{"aliases": ["src/calendarflow/core.py", "core.py"], "kind": "file", "value": "src/calendarflow/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/calendarflow/store.py", "store.py"], "kind": "file", "value": "src/calendarflow/store.py"}, {"...
[{"answers": ["src/calendarflow/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-s1-00550-twin
calendarflow
ctx_compress-s1-00550
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_window lets defaults overwrite explicit overrides. It lives in src/calendarflow/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 calendarflow layout. [tool_use ki...
22,171
1,108
42
81
false
synth:583569567764500037@1
synth
[{"aliases": ["src/calendarflow/core.py", "core.py"], "kind": "file", "value": "src/calendarflow/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/calendarflow/store.py", "store.py"], "kind": "file", "value": "src/calendarflow/store.py"}, {"...
[{"answers": ["src/calendarflow/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-s1-00551
bearingdesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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 raises BearingShapeError for bytes input in strict mode instead of decoding it. It lives in src/bearingdesk/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 ...
91,352
3,072
112
129
false
synth:2259309311973392006@1.56507
synth
[{"aliases": ["src/bearingdesk/rules.py", "rules.py"], "kind": "file", "value": "src/bearingdesk/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/bearingdesk/audit.py", "audit.py"], "kind": "file", "value": "src/bearingdesk/audit.py"}, ...
[{"answers": ["src/bearingdesk/rules.py", "rules.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's mo...
ctx_compress-s1-00551-twin
bearingdesk
ctx_compress-s1-00551
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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 raises BearingShapeError for bytes input in strict mode instead of decoding it. It lives in src/bearingdesk/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 ...
47,145
2,357
64
129
false
synth:2259309311973392006@1.56507
synth
[{"aliases": ["src/bearingdesk/rules.py", "rules.py"], "kind": "file", "value": "src/bearingdesk/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["src/bearingdesk/audit.py", "audit.py"], "kind": "file", "value": "src/bearingdesk/audit.py"}, ...
[{"answers": ["src/bearingdesk/rules.py", "rules.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's mo...
ctx_compress-s1-00552
creditworks
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_window skips the first page; page 1 returns what should be page 2. It lives in src/creditworks/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 creditworks la...
56,080
2,804
78
89
false
synth:4928657484909602766@1.1348
synth
[{"aliases": ["src/creditworks/core.py", "core.py"], "kind": "file", "value": "src/creditworks/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["data/balance_2.csv", "balance_2.csv"], "kind": "file", "value": "data/balance_2.csv"}, {"aliases": [...
[{"answers": ["src/creditworks/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-s1-00552-twin
creditworks
ctx_compress-s1-00552
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_window skips the first page; page 1 returns what should be page 2. It lives in src/creditworks/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 creditworks la...
29,441
1,472
46
89
false
synth:4928657484909602766@1.1348
synth
[{"aliases": ["src/creditworks/core.py", "core.py"], "kind": "file", "value": "src/creditworks/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["data/balance_2.csv", "balance_2.csv"], "kind": "file", "value": "data/balance_2.csv"}, {"aliases": [...
[{"answers": ["src/creditworks/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-s1-00553
footnotekit
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_heading_page averages one value too many once the window is full. It lives in src/footnotekit/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 footnotekit layout. [...
69,046
3,072
90
99
false
synth:2550694142451116406@1.97893
synth
[{"aliases": ["src/footnotekit/io.py", "io.py"], "kind": "file", "value": "src/footnotekit/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["data/heading_1.csv", "heading_1.csv"], "kind": "file", "value": "data/heading_1.csv"}, {"aliases": ["docs/change...
[{"answers": ["src/footnotekit/io.py", "io.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most rec...