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-s1-01306-twin
indexkit
ctx_compress-s1-01306
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] anchor_revision leaves a trailing dash when the title ends in punctuation. It lives in src/indexkit/compat.py; `python -m pytest tests/test_compat.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the indexkit layout....
31,223
1,561
59
116
false
synth:2797295971119267769@1.24585
synth
[{"aliases": ["src/indexkit/compat.py", "compat.py"], "kind": "file", "value": "src/indexkit/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["data/index_0.csv", "index_0.csv"], "kind": "file", "value": "data/index_0.csv"}, {"aliases": [...
[{"answers": ["src/indexkit/compat.py", "compat.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's mos...
ctx_compress-s1-01307
lexemeflow
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_stem_once double counts carried amounts when the period closes on a weekend. It lives in src/lexemeflow/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 lexemeflow ...
74,432
3,072
90
107
false
synth:177193391408286757@2.11583
synth
[{"aliases": ["src/lexemeflow/cli.py", "cli.py"], "kind": "file", "value": "src/lexemeflow/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["src/lexemeflow/tables.py", "tables.py"], "kind": "file", "value": "src/lexemeflow/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-s1-01307-twin
lexemeflow
ctx_compress-s1-01307
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_stem_once double counts carried amounts when the period closes on a weekend. It lives in src/lexemeflow/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 lexemeflow ...
45,643
2,282
55
107
false
synth:177193391408286757@2.11583
synth
[{"aliases": ["src/lexemeflow/cli.py", "cli.py"], "kind": "file", "value": "src/lexemeflow/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["src/lexemeflow/tables.py", "tables.py"], "kind": "file", "value": "src/lexemeflow/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": ["src/lexemeflow/tables.py", "tables.py...
ctx_compress-s1-01308
principaldesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] revoke_session_once leaves a trailing dash when the title ends in punctuation. It lives in src/principaldesk/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 principal...
79,596
3,072
106
125
false
synth:7968626188701107252@1.96598
synth
[{"aliases": ["src/principaldesk/store.py", "store.py"], "kind": "file", "value": "src/principaldesk/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["docs/changelog.md", "changelog.md"], "kind": "file", "value": "docs/CHANGELOG.md"}, {"alia...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["core.py", "src/principaldesk/core.py"...
ctx_compress-s1-01308-twin
principaldesk
ctx_compress-s1-01308
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] revoke_session_once leaves a trailing dash when the title ends in punctuation. It lives in src/principaldesk/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 principal...
39,138
1,956
57
125
false
synth:7968626188701107252@1.96598
synth
[{"aliases": ["src/principaldesk/store.py", "store.py"], "kind": "file", "value": "src/principaldesk/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["docs/changelog.md", "changelog.md"], "kind": "file", "value": "docs/CHANGELOG.md"}, {"alia...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["data/session_6.csv", "session_6.csv"]...
ctx_compress-s1-01309
cueflow
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] truncate_snippet double counts carried amounts when the period closes on a weekend. It lives in src/cueflow/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 cueflow ...
55,149
2,757
90
105
false
synth:1278383677565697018@1.38072
synth
[{"aliases": ["src/cueflow/compat.py", "compat.py"], "kind": "file", "value": "src/cueflow/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["src/cueflow/audit.py", "audit.py"], "kind": "file", "value": "src/cueflow/audit.py"}, {"aliases"...
[{"answers": ["src/cueflow/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 most...
ctx_compress-s1-01309-twin
cueflow
ctx_compress-s1-01309
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] truncate_snippet double counts carried amounts when the period closes on a weekend. It lives in src/cueflow/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 cueflow ...
35,365
1,768
62
105
false
synth:1278383677565697018@1.38072
synth
[{"aliases": ["src/cueflow/compat.py", "compat.py"], "kind": "file", "value": "src/cueflow/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["src/cueflow/audit.py", "audit.py"], "kind": "file", "value": "src/cueflow/audit.py"}, {"aliases"...
[{"answers": ["src/cueflow/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 most...
ctx_compress-s1-01310
glyphflow
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] truncate_slug raises ValueError when a value itself contains an equals sign. It lives in src/glyphflow/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 glyphflow layout....
55,211
2,760
87
102
false
synth:3760749797320229851@1.55808
synth
[{"aliases": ["src/glyphflow/core.py", "core.py"], "kind": "file", "value": "src/glyphflow/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/glyphflow/io.py", "io.py"], "kind": "file", "value": "src/glyphflow/io.py"}, {"aliases": ["data/cue_...
[{"answers": ["src/glyphflow/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-s1-01310-twin
glyphflow
ctx_compress-s1-01310
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] truncate_slug raises ValueError when a value itself contains an equals sign. It lives in src/glyphflow/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 glyphflow layout....
30,419
1,520
49
102
false
synth:3760749797320229851@1.55808
synth
[{"aliases": ["src/glyphflow/core.py", "core.py"], "kind": "file", "value": "src/glyphflow/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/glyphflow/io.py", "io.py"], "kind": "file", "value": "src/glyphflow/io.py"}, {"aliases": ["data/cue_...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["test_io.py", "tests/test_io.py"], "ki...
ctx_compress-s1-01311
glyphforge
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] truncate_glyph ignores the cap, so late attempts wait for minutes. It lives in src/glyphforge/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 glyphforge layout. [tool...
103,492
3,072
105
132
false
synth:2617035609121624213@2.02228
synth
[{"aliases": ["src/glyphforge/rules.py", "rules.py"], "kind": "file", "value": "src/glyphforge/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["data/chunk_0.csv", "chunk_0.csv"], "kind": "file", "value": "data/chunk_0.csv"}, {"aliases": ["d...
[{"answers": ["src/glyphforge/rules.py", "rules.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's mos...
ctx_compress-s1-01311-twin
glyphforge
ctx_compress-s1-01311
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] truncate_glyph ignores the cap, so late attempts wait for minutes. It lives in src/glyphforge/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 glyphforge layout. [tool...
55,472
2,773
63
132
false
synth:2617035609121624213@2.02228
synth
[{"aliases": ["src/glyphforge/rules.py", "rules.py"], "kind": "file", "value": "src/glyphforge/rules.py"}, {"aliases": ["tests/test_rules.py", "test_rules.py"], "kind": "file", "value": "tests/test_rules.py"}, {"aliases": ["data/chunk_0.csv", "chunk_0.csv"], "kind": "file", "value": "data/chunk_0.csv"}, {"aliases": ["d...
[{"answers": ["src/glyphforge/rules.py", "rules.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's mos...
ctx_compress-s1-01312
glyphdesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] tokenise_glyph raises ChunkBoundError for bytes input in strict mode instead of decoding it. It lives in src/glyphdesk/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 t...
113,155
3,072
123
142
false
synth:7447567892287765411@2.3373
synth
[{"aliases": ["src/glyphdesk/format.py", "format.py"], "kind": "file", "value": "src/glyphdesk/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["src/glyphdesk/cli.py", "cli.py"], "kind": "file", "value": "src/glyphdesk/cli.py"}, {"aliase...
[{"answers": ["src/glyphdesk/format.py", "format.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's mo...
ctx_compress-s1-01312-twin
glyphdesk
ctx_compress-s1-01312
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_glyph raises ChunkBoundError for bytes input in strict mode instead of decoding it. It lives in src/glyphdesk/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 t...
75,188
3,072
75
142
false
synth:7447567892287765411@2.3373
synth
[{"aliases": ["src/glyphdesk/format.py", "format.py"], "kind": "file", "value": "src/glyphdesk/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["src/glyphdesk/cli.py", "cli.py"], "kind": "file", "value": "src/glyphdesk/cli.py"}, {"aliase...
[{"answers": ["src/glyphdesk/format.py", "format.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's mo...
ctx_compress-s1-01313
reorderworks
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] count_stock raises ValueError when a value itself contains an equals sign. It lives in src/reorderworks/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 reorderworks...
24,307
1,215
46
57
false
synth:5330979866543372346@1
synth
[{"aliases": ["src/reorderworks/format.py", "format.py"], "kind": "file", "value": "src/reorderworks/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["data/pallet_3.csv", "pallet_3.csv"], "kind": "file", "value": "data/pallet_3.csv"}, {"...
[{"answers": ["src/reorderworks/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-s1-01313-twin
reorderworks
ctx_compress-s1-01313
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] count_stock raises ValueError when a value itself contains an equals sign. It lives in src/reorderworks/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 reorderworks...
18,745
1,000
32
57
false
synth:5330979866543372346@1
synth
[{"aliases": ["src/reorderworks/format.py", "format.py"], "kind": "file", "value": "src/reorderworks/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["data/pallet_3.csv", "pallet_3.csv"], "kind": "file", "value": "data/pallet_3.csv"}, {"...
[{"answers": ["src/reorderworks/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-s1-01314
fixworks
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] cluster_waypoint_page misses keys that differ only in case. It lives in src/fixworks/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 fixworks layout. [tool_use kind...
104,782
3,072
118
127
false
synth:912974505057371321@1.95178
synth
[{"aliases": ["src/fixworks/limits.py", "limits.py"], "kind": "file", "value": "src/fixworks/limits.py"}, {"aliases": ["tests/test_limits.py", "test_limits.py"], "kind": "file", "value": "tests/test_limits.py"}, {"aliases": ["data/zone_9.csv", "zone_9.csv"], "kind": "file", "value": "data/zone_9.csv"}, {"aliases": ["sr...
[{"answers": ["src/fixworks/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 mos...
ctx_compress-s1-01314-twin
fixworks
ctx_compress-s1-01314
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] cluster_waypoint_page misses keys that differ only in case. It lives in src/fixworks/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 fixworks layout. [tool_use kind...
59,544
2,977
76
127
false
synth:912974505057371321@1.95178
synth
[{"aliases": ["src/fixworks/limits.py", "limits.py"], "kind": "file", "value": "src/fixworks/limits.py"}, {"aliases": ["tests/test_limits.py", "test_limits.py"], "kind": "file", "value": "tests/test_limits.py"}, {"aliases": ["data/zone_9.csv", "zone_9.csv"], "kind": "file", "value": "data/zone_9.csv"}, {"aliases": ["sr...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["test_format.py", "tests/test_format.p...
ctx_compress-s1-01315
rosterledger
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] assign_shift leaves a trailing dash when the title ends in punctuation. It lives in src/rosterledger/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 rosterledger la...
51,160
2,558
74
81
false
synth:6904667813462595726@1
synth
[{"aliases": ["src/rosterledger/format.py", "format.py"], "kind": "file", "value": "src/rosterledger/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/rosterledger/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-s1-01315-twin
rosterledger
ctx_compress-s1-01315
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] assign_shift leaves a trailing dash when the title ends in punctuation. It lives in src/rosterledger/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 rosterledger la...
30,993
1,549
48
81
false
synth:6904667813462595726@1
synth
[{"aliases": ["src/rosterledger/format.py", "format.py"], "kind": "file", "value": "src/rosterledger/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": ["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/rosterledger/schema....
ctx_compress-s1-01316
revisiondesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_batch raises ValueError when a value itself contains an equals sign. It lives in src/revisiondesk/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 re...
44,769
2,238
89
96
false
synth:5173760834227810740@1
synth
[{"aliases": ["src/revisiondesk/report.py", "report.py"], "kind": "file", "value": "src/revisiondesk/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["src/revisiondesk/io.py", "io.py"], "kind": "file", "value": "src/revisiondesk/io.py"},...
[{"answers": ["src/revisiondesk/report.py", "report.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's...
ctx_compress-s1-01316-twin
revisiondesk
ctx_compress-s1-01316
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_batch raises ValueError when a value itself contains an equals sign. It lives in src/revisiondesk/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 re...
27,657
1,382
55
96
false
synth:5173760834227810740@1
synth
[{"aliases": ["src/revisiondesk/report.py", "report.py"], "kind": "file", "value": "src/revisiondesk/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["src/revisiondesk/io.py", "io.py"], "kind": "file", "value": "src/revisiondesk/io.py"},...
[{"answers": ["src/revisiondesk/report.py", "report.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's...
ctx_compress-s1-01317
snippetledger
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_caption returns the upper bound for every value instead of clamping into the range. It lives in src/snippetledger/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 t...
51,495
2,574
93
108
false
synth:6579500893554068122@1.36787
synth
[{"aliases": ["src/snippetledger/audit.py", "audit.py"], "kind": "file", "value": "src/snippetledger/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["src/snippetledger/core.py", "core.py"], "kind": "file", "value": "src/snippetledger/core.p...
[{"answers": ["src/snippetledger/audit.py", "audit.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's ...
ctx_compress-s1-01317-twin
snippetledger
ctx_compress-s1-01317
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_caption returns the upper bound for every value instead of clamping into the range. It lives in src/snippetledger/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 t...
31,630
1,581
53
108
false
synth:6579500893554068122@1.36787
synth
[{"aliases": ["src/snippetledger/audit.py", "audit.py"], "kind": "file", "value": "src/snippetledger/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["src/snippetledger/core.py", "core.py"], "kind": "file", "value": "src/snippetledger/core.p...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["audit.py", "src/snippetledger/audit.p...
ctx_compress-s1-01318
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_tile_once returns one item too few on every page. It lives in src/bearingdesk/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 bearingdesk layout. [tool_use kind...
72,578
3,072
109
124
false
synth:3588061840218893137@1.51485
synth
[{"aliases": ["src/bearingdesk/core.py", "core.py"], "kind": "file", "value": "src/bearingdesk/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/bearingdesk/tables.py", "tables.py"], "kind": "file", "value": "src/bearingdesk/tables.py"}, {"a...
[{"answers": ["src/bearingdesk/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-s1-01318-twin
bearingdesk
ctx_compress-s1-01318
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_tile_once returns one item too few on every page. It lives in src/bearingdesk/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 bearingdesk layout. [tool_use kind...
47,174
2,358
67
124
false
synth:3588061840218893137@1.51485
synth
[{"aliases": ["src/bearingdesk/core.py", "core.py"], "kind": "file", "value": "src/bearingdesk/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/bearingdesk/tables.py", "tables.py"], "kind": "file", "value": "src/bearingdesk/tables.py"}, {"a...
[{"answers": ["src/bearingdesk/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-s1-01319
shipmentflow
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] count_sku raises ValueError when a value itself contains an equals sign. It lives in src/shipmentflow/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 shipmentflow l...
26,805
1,340
56
65
false
synth:8285837295317792840@1
synth
[{"aliases": ["src/shipmentflow/compat.py", "compat.py"], "kind": "file", "value": "src/shipmentflow/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["docs/changelog.md", "changelog.md"], "kind": "file", "value": "docs/CHANGELOG.md"}, {"...
[{"answers": ["src/shipmentflow/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-s1-01319-twin
shipmentflow
ctx_compress-s1-01319
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] count_sku raises ValueError when a value itself contains an equals sign. It lives in src/shipmentflow/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 shipmentflow l...
11,926
1,000
37
65
false
synth:8285837295317792840@1
synth
[{"aliases": ["src/shipmentflow/compat.py", "compat.py"], "kind": "file", "value": "src/shipmentflow/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["docs/changelog.md", "changelog.md"], "kind": "file", "value": "docs/CHANGELOG.md"}, {"...
[{"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/shipmentflow/tables.py", "tables....
ctx_compress-s1-01320
footnotedesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_heading keeps leading spaces in keys so lookups after a semicolon fail. It lives in src/footnotedesk/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 footnotedesk...
94,185
3,072
120
137
false
synth:1920768351515217701@2.06335
synth
[{"aliases": ["src/footnotedesk/audit.py", "audit.py"], "kind": "file", "value": "src/footnotedesk/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["src/footnotedesk/cli.py", "cli.py"], "kind": "file", "value": "src/footnotedesk/cli.py"}, {"...
[{"answers": ["src/footnotedesk/audit.py", "audit.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's m...
ctx_compress-s1-01320-twin
footnotedesk
ctx_compress-s1-01320
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_heading keeps leading spaces in keys so lookups after a semicolon fail. It lives in src/footnotedesk/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 footnotedesk...
53,467
2,673
74
137
false
synth:1920768351515217701@2.06335
synth
[{"aliases": ["src/footnotedesk/audit.py", "audit.py"], "kind": "file", "value": "src/footnotedesk/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["src/footnotedesk/cli.py", "cli.py"], "kind": "file", "value": "src/footnotedesk/cli.py"}, {"...
[{"answers": ["src/footnotedesk/audit.py", "audit.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's m...
ctx_compress-s1-01321
rosterdesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_calendar_safe keeps leading spaces in keys so lookups after a semicolon fail. It lives in src/rosterdesk/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 ros...
109,853
3,072
115
134
false
synth:1892704477675559056@2.26715
synth
[{"aliases": ["src/rosterdesk/report.py", "report.py"], "kind": "file", "value": "src/rosterdesk/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["src/rosterdesk/schema.py", "schema.py"], "kind": "file", "value": "src/rosterdesk/schema.p...
[{"answers": ["src/rosterdesk/report.py", "report.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's m...
ctx_compress-s1-01321-twin
rosterdesk
ctx_compress-s1-01321
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_calendar_safe keeps leading spaces in keys so lookups after a semicolon fail. It lives in src/rosterdesk/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 ros...
73,291
3,072
77
134
false
synth:1892704477675559056@2.26715
synth
[{"aliases": ["src/rosterdesk/report.py", "report.py"], "kind": "file", "value": "src/rosterdesk/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["src/rosterdesk/schema.py", "schema.py"], "kind": "file", "value": "src/rosterdesk/schema.p...
[{"answers": ["src/rosterdesk/report.py", "report.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's m...
ctx_compress-s1-01322
tokenledger
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] narrow_nonce lets defaults overwrite explicit overrides. It lives in src/tokenledger/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 tokenledger layout. [tool_use k...
33,343
1,667
68
83
false
synth:8233955398042875853@1
synth
[{"aliases": ["src/tokenledger/limits.py", "limits.py"], "kind": "file", "value": "src/tokenledger/limits.py"}, {"aliases": ["tests/test_limits.py", "test_limits.py"], "kind": "file", "value": "tests/test_limits.py"}, {"aliases": ["src/tokenledger/report.py", "report.py"], "kind": "file", "value": "src/tokenledger/repo...
[{"answers": ["src/tokenledger/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-s1-01322-twin
tokenledger
ctx_compress-s1-01322
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] narrow_nonce lets defaults overwrite explicit overrides. It lives in src/tokenledger/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 tokenledger layout. [tool_use k...
17,460
1,000
40
83
false
synth:8233955398042875853@1
synth
[{"aliases": ["src/tokenledger/limits.py", "limits.py"], "kind": "file", "value": "src/tokenledger/limits.py"}, {"aliases": ["tests/test_limits.py", "test_limits.py"], "kind": "file", "value": "tests/test_limits.py"}, {"aliases": ["src/tokenledger/report.py", "report.py"], "kind": "file", "value": "src/tokenledger/repo...
[{"answers": ["src/tokenledger/limits.py", "limits.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's ...
ctx_compress-s1-01323
slugledger
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] tokenise_snippet raises ValueError when a value itself contains an equals sign. It lives in src/slugledger/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 slugledge...
79,262
3,072
99
122
false
synth:3315891896602410155@1.69645
synth
[{"aliases": ["src/slugledger/compat.py", "compat.py"], "kind": "file", "value": "src/slugledger/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["src/slugledger/limits.py", "limits.py"], "kind": "file", "value": "src/slugledger/limits.p...
[{"answers": ["src/slugledger/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-s1-01323-twin
slugledger
ctx_compress-s1-01323
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_snippet raises ValueError when a value itself contains an equals sign. It lives in src/slugledger/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 slugledge...
40,856
2,042
59
122
false
synth:3315891896602410155@1.69645
synth
[{"aliases": ["src/slugledger/compat.py", "compat.py"], "kind": "file", "value": "src/slugledger/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["src/slugledger/limits.py", "limits.py"], "kind": "file", "value": "src/slugledger/limits.p...
[{"answers": ["src/slugledger/compat.py", "compat.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's m...
ctx_compress-s1-01324
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_draft raises ValueError when a value itself contains an equals sign. It lives in src/headingledger/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 headingledge...
54,190
2,709
96
113
false
synth:7770084829414882909@1.42425
synth
[{"aliases": ["src/headingledger/store.py", "store.py"], "kind": "file", "value": "src/headingledger/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/headingledger/rules.py", "rules.py"], "kind": "file", "value": "src/headingledger/rule...
[{"answers": ["src/headingledger/store.py", "store.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's ...
ctx_compress-s1-01324-twin
headingledger
ctx_compress-s1-01324
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_draft raises ValueError when a value itself contains an equals sign. It lives in src/headingledger/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 headingledge...
29,234
1,461
57
113
false
synth:7770084829414882909@1.42425
synth
[{"aliases": ["src/headingledger/store.py", "store.py"], "kind": "file", "value": "src/headingledger/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/headingledger/rules.py", "rules.py"], "kind": "file", "value": "src/headingledger/rule...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["test_format.py", "tests/test_format.p...
ctx_compress-s1-01325
creditdesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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 rejects a range where lo equals hi, which is a valid single point range. It lives in src/creditdesk/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 cred...
35,385
1,769
70
89
false
synth:828321341660598610@1
synth
[{"aliases": ["src/creditdesk/audit.py", "audit.py"], "kind": "file", "value": "src/creditdesk/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["src/creditdesk/core.py", "core.py"], "kind": "file", "value": "src/creditdesk/core.py"}, {"alias...
[{"answers": ["src/creditdesk/audit.py", "audit.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's mos...
ctx_compress-s1-01325-twin
creditdesk
ctx_compress-s1-01325
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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 rejects a range where lo equals hi, which is a valid single point range. It lives in src/creditdesk/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 cred...
22,942
1,147
44
89
false
synth:828321341660598610@1
synth
[{"aliases": ["src/creditdesk/audit.py", "audit.py"], "kind": "file", "value": "src/creditdesk/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["src/creditdesk/core.py", "core.py"], "kind": "file", "value": "src/creditdesk/core.py"}, {"alias...
[{"answers": ["src/creditdesk/audit.py", "audit.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's mos...
ctx_compress-s1-01326
rosterworks
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_strict misses keys that differ only in case. It lives in src/rosterworks/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 rosterworks layout. [tool_use kin...
42,230
2,111
86
95
false
synth:7696209202036470936@1
synth
[{"aliases": ["src/rosterworks/core.py", "core.py"], "kind": "file", "value": "src/rosterworks/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/rosterworks/audit.py", "audit.py"], "kind": "file", "value": "src/rosterworks/audit.py"}, {"alia...
[{"answers": ["src/rosterworks/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-s1-01326-twin
rosterworks
ctx_compress-s1-01326
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_strict misses keys that differ only in case. It lives in src/rosterworks/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 rosterworks layout. [tool_use kin...
21,095
1,054
50
95
false
synth:7696209202036470936@1
synth
[{"aliases": ["src/rosterworks/core.py", "core.py"], "kind": "file", "value": "src/rosterworks/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/rosterworks/audit.py", "audit.py"], "kind": "file", "value": "src/rosterworks/audit.py"}, {"alia...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["audit.py", "src/rosterworks/audit.py"...
ctx_compress-s1-01327
cadencedesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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 misses keys that differ only in case. It lives in src/cadencedesk/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 cadencedesk layout. [tool_use kind=search ...
85,491
3,072
98
107
false
synth:6213384905674735571@1.60745
synth
[{"aliases": ["src/cadencedesk/cli.py", "cli.py"], "kind": "file", "value": "src/cadencedesk/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["src/cadencedesk/limits.py", "limits.py"], "kind": "file", "value": "src/cadencedesk/limits.py"}, {"aliases...
[{"answers": ["src/cadencedesk/cli.py", "cli.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most r...
ctx_compress-s1-01327-twin
cadencedesk
ctx_compress-s1-01327
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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 misses keys that differ only in case. It lives in src/cadencedesk/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 cadencedesk layout. [tool_use kind=search ...
53,811
2,690
64
107
false
synth:6213384905674735571@1.60745
synth
[{"aliases": ["src/cadencedesk/cli.py", "cli.py"], "kind": "file", "value": "src/cadencedesk/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["src/cadencedesk/limits.py", "limits.py"], "kind": "file", "value": "src/cadencedesk/limits.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/roster_1.csv", "roster_1.csv"], ...
ctx_compress-s1-01328
sectionflow
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] fold_section_all leaves a trailing dash when the title ends in punctuation. It lives in src/sectionflow/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 sectionflow ...
61,823
3,072
92
113
false
synth:760143527984317644@1.2424
synth
[{"aliases": ["src/sectionflow/limits.py", "limits.py"], "kind": "file", "value": "src/sectionflow/limits.py"}, {"aliases": ["tests/test_limits.py", "test_limits.py"], "kind": "file", "value": "tests/test_limits.py"}, {"aliases": ["data/index_0.csv", "index_0.csv"], "kind": "file", "value": "data/index_0.csv"}, {"alias...
[{"answers": ["src/sectionflow/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-s1-01328-twin
sectionflow
ctx_compress-s1-01328
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_section_all leaves a trailing dash when the title ends in punctuation. It lives in src/sectionflow/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 sectionflow ...
30,771
1,538
51
113
false
synth:760143527984317644@1.2424
synth
[{"aliases": ["src/sectionflow/limits.py", "limits.py"], "kind": "file", "value": "src/sectionflow/limits.py"}, {"aliases": ["tests/test_limits.py", "test_limits.py"], "kind": "file", "value": "tests/test_limits.py"}, {"aliases": ["data/index_0.csv", "index_0.csv"], "kind": "file", "value": "data/index_0.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": ["limits.py", "src/sectionflow/limits.p...
ctx_compress-s1-01329
tokenforge
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_batch drops override values of 0, False and empty string. It lives in src/tokenforge/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 tokenforge layout....
56,178
2,808
98
119
false
synth:8877045387777085570@1.29125
synth
[{"aliases": ["src/tokenforge/format.py", "format.py"], "kind": "file", "value": "src/tokenforge/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"}, {"alia...
[{"answers": ["src/tokenforge/format.py", "format.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's m...
ctx_compress-s1-01329-twin
tokenforge
ctx_compress-s1-01329
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_batch drops override values of 0, False and empty string. It lives in src/tokenforge/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 tokenforge layout....
32,246
1,612
55
119
false
synth:8877045387777085570@1.29125
synth
[{"aliases": ["src/tokenforge/format.py", "format.py"], "kind": "file", "value": "src/tokenforge/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"}, {"alia...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["test_cli.py", "tests/test_cli.py"], "...
ctx_compress-s1-01330
rosterkit
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_slot_batch raises BookingStateError for bytes input in strict mode instead of decoding it. It lives in src/rosterkit/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...
19,573
1,000
48
59
false
synth:5360052414058052113@1
synth
[{"aliases": ["src/rosterkit/core.py", "core.py"], "kind": "file", "value": "src/rosterkit/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/rosterkit/schema.py", "schema.py"], "kind": "file", "value": "src/rosterkit/schema.py"}, {"aliases":...
[{"answers": ["src/rosterkit/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-s1-01330-twin
rosterkit
ctx_compress-s1-01330
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_slot_batch raises BookingStateError for bytes input in strict mode instead of decoding it. It lives in src/rosterkit/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...
14,829
1,000
34
59
false
synth:5360052414058052113@1
synth
[{"aliases": ["src/rosterkit/core.py", "core.py"], "kind": "file", "value": "src/rosterkit/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/rosterkit/schema.py", "schema.py"], "kind": "file", "value": "src/rosterkit/schema.py"}, {"aliases":...
[{"answers": ["src/rosterkit/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-s1-01331
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] paginate_draft raises IndexStateError for bytes input in strict mode instead of decoding it. It lives in src/footnotekit/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...
52,819
2,640
92
107
false
synth:2701428462261319304@1.26555
synth
[{"aliases": ["src/footnotekit/core.py", "core.py"], "kind": "file", "value": "src/footnotekit/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["docs/changelog.md", "changelog.md"], "kind": "file", "value": "docs/CHANGELOG.md"}, {"aliases": ["sr...
[{"answers": ["src/footnotekit/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-s1-01331-twin
footnotekit
ctx_compress-s1-01331
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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 raises IndexStateError for bytes input in strict mode instead of decoding it. It lives in src/footnotekit/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...
33,035
1,651
62
107
false
synth:2701428462261319304@1.26555
synth
[{"aliases": ["src/footnotekit/core.py", "core.py"], "kind": "file", "value": "src/footnotekit/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["docs/changelog.md", "changelog.md"], "kind": "file", "value": "docs/CHANGELOG.md"}, {"aliases": ["sr...
[{"answers": ["src/footnotekit/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-s1-01332
policyforge
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] narrow_principal keeps leading spaces in keys so lookups after a semicolon fail. It lives in src/policyforge/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 policyfor...
34,401
1,720
56
69
false
synth:5979060720554632073@1
synth
[{"aliases": ["src/policyforge/audit.py", "audit.py"], "kind": "file", "value": "src/policyforge/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["src/policyforge/report.py", "report.py"], "kind": "file", "value": "src/policyforge/report.py"...
[{"answers": ["src/policyforge/audit.py", "audit.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's mo...
ctx_compress-s1-01332-twin
policyforge
ctx_compress-s1-01332
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] narrow_principal keeps leading spaces in keys so lookups after a semicolon fail. It lives in src/policyforge/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 policyfor...
17,661
1,000
36
69
false
synth:5979060720554632073@1
synth
[{"aliases": ["src/policyforge/audit.py", "audit.py"], "kind": "file", "value": "src/policyforge/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["src/policyforge/report.py", "report.py"], "kind": "file", "value": "src/policyforge/report.py"...
[{"answers": ["src/policyforge/audit.py", "audit.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's mo...
ctx_compress-s1-01333
seriesworks
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_bucket_batch drops override values of 0, False and empty string. It lives in src/seriesworks/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 seriesworks layo...
94,629
3,072
108
125
false
synth:5136292464237998853@1.9064
synth
[{"aliases": ["src/seriesworks/limits.py", "limits.py"], "kind": "file", "value": "src/seriesworks/limits.py"}, {"aliases": ["tests/test_limits.py", "test_limits.py"], "kind": "file", "value": "tests/test_limits.py"}, {"aliases": ["data/bucket_8.csv", "bucket_8.csv"], "kind": "file", "value": "data/bucket_8.csv"}, {"al...
[{"answers": ["src/seriesworks/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-s1-01333-twin
seriesworks
ctx_compress-s1-01333
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_bucket_batch drops override values of 0, False and empty string. It lives in src/seriesworks/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 seriesworks layo...
62,698
3,072
70
125
false
synth:5136292464237998853@1.9064
synth
[{"aliases": ["src/seriesworks/limits.py", "limits.py"], "kind": "file", "value": "src/seriesworks/limits.py"}, {"aliases": ["tests/test_limits.py", "test_limits.py"], "kind": "file", "value": "tests/test_limits.py"}, {"aliases": ["data/bucket_8.csv", "bucket_8.csv"], "kind": "file", "value": "data/bucket_8.csv"}, {"al...
[{"answers": ["src/seriesworks/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-s1-01334
stockforge
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] age_pallet_page double counts carried amounts when the period closes on a weekend. It lives in src/stockforge/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 stockf...
50,673
2,533
106
121
false
synth:2397294486536975002@1.14145
synth
[{"aliases": ["src/stockforge/format.py", "format.py"], "kind": "file", "value": "src/stockforge/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["data/sku_2.csv", "sku_2.csv"], "kind": "file", "value": "data/sku_2.csv"}, {"aliases": ["d...
[{"answers": ["src/stockforge/format.py", "format.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's m...
ctx_compress-s1-01334-twin
stockforge
ctx_compress-s1-01334
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] age_pallet_page double counts carried amounts when the period closes on a weekend. It lives in src/stockforge/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 stockf...
30,602
1,530
55
121
false
synth:2397294486536975002@1.14145
synth
[{"aliases": ["src/stockforge/format.py", "format.py"], "kind": "file", "value": "src/stockforge/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["data/sku_2.csv", "sku_2.csv"], "kind": "file", "value": "data/sku_2.csv"}, {"aliases": ["d...
[{"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/stockforge/format.py...
ctx_compress-s1-01335
receiptworks
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_credit_once double counts carried amounts when the period closes on a weekend. It lives in src/receiptworks/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 rece...
55,740
2,787
93
100
false
synth:5481042449379731946@1
synth
[{"aliases": ["src/receiptworks/limits.py", "limits.py"], "kind": "file", "value": "src/receiptworks/limits.py"}, {"aliases": ["tests/test_limits.py", "test_limits.py"], "kind": "file", "value": "tests/test_limits.py"}, {"aliases": ["data/credit_3.csv", "credit_3.csv"], "kind": "file", "value": "data/credit_3.csv"}, {"...
[{"answers": ["src/receiptworks/limits.py", "limits.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's...
ctx_compress-s1-01335-twin
receiptworks
ctx_compress-s1-01335
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_credit_once double counts carried amounts when the period closes on a weekend. It lives in src/receiptworks/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 rece...
32,023
1,601
49
100
false
synth:5481042449379731946@1
synth
[{"aliases": ["src/receiptworks/limits.py", "limits.py"], "kind": "file", "value": "src/receiptworks/limits.py"}, {"aliases": ["tests/test_limits.py", "test_limits.py"], "kind": "file", "value": "tests/test_limits.py"}, {"aliases": ["data/credit_3.csv", "credit_3.csv"], "kind": "file", "value": "data/credit_3.csv"}, {"...
[{"answers": ["src/receiptworks/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-s1-01336
headingworks
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_section returns the upper bound for every value instead of clamping into the range. It lives in src/headingworks/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 ...
57,570
2,878
107
116
false
synth:2343199744597986666@1.26165
synth
[{"aliases": ["src/headingworks/core.py", "core.py"], "kind": "file", "value": "src/headingworks/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/headingworks/compat.py", "compat.py"], "kind": "file", "value": "src/headingworks/compat.py"},...
[{"answers": ["src/headingworks/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-s1-01336-twin
headingworks
ctx_compress-s1-01336
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_section returns the upper bound for every value instead of clamping into the range. It lives in src/headingworks/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 ...
30,091
1,504
63
116
false
synth:2343199744597986666@1.26165
synth
[{"aliases": ["src/headingworks/core.py", "core.py"], "kind": "file", "value": "src/headingworks/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["src/headingworks/compat.py", "compat.py"], "kind": "file", "value": "src/headingworks/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": ["data/section_0.csv", "section_0.csv"]...
ctx_compress-s1-01337
draftledger
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_index_page misses keys that differ only in case. It lives in src/draftledger/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 draftledger layout. [tool_use kind...
53,417
2,670
71
80
false
synth:6706865837727901685@1
synth
[{"aliases": ["src/draftledger/store.py", "store.py"], "kind": "file", "value": "src/draftledger/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["data/section_0.csv", "section_0.csv"], "kind": "file", "value": "data/section_0.csv"}, {"alias...
[{"answers": ["src/draftledger/store.py", "store.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's mo...
ctx_compress-s1-01337-twin
draftledger
ctx_compress-s1-01337
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_index_page misses keys that differ only in case. It lives in src/draftledger/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 draftledger layout. [tool_use kind...
24,418
1,220
37
80
false
synth:6706865837727901685@1
synth
[{"aliases": ["src/draftledger/store.py", "store.py"], "kind": "file", "value": "src/draftledger/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["data/section_0.csv", "section_0.csv"], "kind": "file", "value": "data/section_0.csv"}, {"alias...
[{"answers": ["src/draftledger/store.py", "store.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's mo...
ctx_compress-s1-01338
creditkit
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] roll_refund returns one item too few on every page. It lives in src/creditkit/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 creditkit layout. [tool_use kind=search name=B...
68,775
3,072
107
124
false
synth:5571374366059000067@1.73155
synth
[{"aliases": ["src/creditkit/io.py", "io.py"], "kind": "file", "value": "src/creditkit/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["data/invoice_2.csv", "invoice_2.csv"], "kind": "file", "value": "data/invoice_2.csv"}, {"aliases": ["src/creditkit/s...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["limits.py", "src/creditkit/limits.py"...
ctx_compress-s1-01338-twin
creditkit
ctx_compress-s1-01338
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] roll_refund returns one item too few on every page. It lives in src/creditkit/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 creditkit layout. [tool_use kind=search name=B...
36,498
1,824
59
124
false
synth:5571374366059000067@1.73155
synth
[{"aliases": ["src/creditkit/io.py", "io.py"], "kind": "file", "value": "src/creditkit/io.py"}, {"aliases": ["tests/test_io.py", "test_io.py"], "kind": "file", "value": "tests/test_io.py"}, {"aliases": ["data/invoice_2.csv", "invoice_2.csv"], "kind": "file", "value": "data/invoice_2.csv"}, {"aliases": ["src/creditkit/s...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["credit_3.csv", "data/credit_3.csv"], ...
ctx_compress-s1-01339
draftkit
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] fold_footnote double counts carried amounts when the period closes on a weekend. It lives in src/draftkit/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 draftkit layout....
76,064
3,072
93
110
false
synth:4023207854664876968@1.82745
synth
[{"aliases": ["src/draftkit/cli.py", "cli.py"], "kind": "file", "value": "src/draftkit/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/draft_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": ["data/margin_5.csv", "margin_5.csv"], ...
ctx_compress-s1-01339-twin
draftkit
ctx_compress-s1-01339
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] fold_footnote double counts carried amounts when the period closes on a weekend. It lives in src/draftkit/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 draftkit layout....
42,830
2,141
53
110
false
synth:4023207854664876968@1.82745
synth
[{"aliases": ["src/draftkit/cli.py", "cli.py"], "kind": "file", "value": "src/draftkit/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/draft_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": ["compat.py", "src/draftkit/compat.py"]...
ctx_compress-s1-01340
spankit
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] bin_trace double counts carried amounts when the period closes on a weekend. It lives in src/spankit/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 spankit layout....
103,471
3,072
116
131
false
synth:477296526672194074@1.8646
synth
[{"aliases": ["src/spankit/compat.py", "compat.py"], "kind": "file", "value": "src/spankit/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["data/counter_3.csv", "counter_3.csv"], "kind": "file", "value": "data/counter_3.csv"}, {"aliases...
[{"answers": ["src/spankit/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 most...
ctx_compress-s1-01340-twin
spankit
ctx_compress-s1-01340
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] bin_trace double counts carried amounts when the period closes on a weekend. It lives in src/spankit/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 spankit layout....
53,715
2,685
64
131
false
synth:477296526672194074@1.8646
synth
[{"aliases": ["src/spankit/compat.py", "compat.py"], "kind": "file", "value": "src/spankit/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["data/counter_3.csv", "counter_3.csv"], "kind": "file", "value": "data/counter_3.csv"}, {"aliases...
[{"answers": ["src/spankit/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 most...
ctx_compress-s1-01341
warehouseworks
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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 rejects a range where lo equals hi, which is a valid single point range. It lives in src/warehouseworks/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 w...
91,166
3,072
118
151
false
synth:2760843583284182696@2.12795
synth
[{"aliases": ["src/warehouseworks/limits.py", "limits.py"], "kind": "file", "value": "src/warehouseworks/limits.py"}, {"aliases": ["tests/test_limits.py", "test_limits.py"], "kind": "file", "value": "tests/test_limits.py"}, {"aliases": ["data/shipment_11.csv", "shipment_11.csv"], "kind": "file", "value": "data/shipment...
[{"answers": ["src/warehouseworks/limits.py", "limits.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent...
ctx_compress-s1-01341-twin
warehouseworks
ctx_compress-s1-01341
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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 rejects a range where lo equals hi, which is a valid single point range. It lives in src/warehouseworks/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 w...
59,762
2,988
80
151
false
synth:2760843583284182696@2.12795
synth
[{"aliases": ["src/warehouseworks/limits.py", "limits.py"], "kind": "file", "value": "src/warehouseworks/limits.py"}, {"aliases": ["tests/test_limits.py", "test_limits.py"], "kind": "file", "value": "tests/test_limits.py"}, {"aliases": ["data/shipment_11.csv", "shipment_11.csv"], "kind": "file", "value": "data/shipment...
[{"answers": ["src/warehouseworks/limits.py", "limits.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent...
ctx_compress-s1-01342
binledger
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] restock_bin returns one item too few on every page. It lives in src/binledger/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 binledger layout. [tool_use kind=searc...
76,804
3,072
98
119
false
synth:8379501523015905309@2.33228
synth
[{"aliases": ["src/binledger/format.py", "format.py"], "kind": "file", "value": "src/binledger/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["src/binledger/cli.py", "cli.py"], "kind": "file", "value": "src/binledger/cli.py"}, {"aliase...
[{"answers": ["src/binledger/format.py", "format.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's mo...
ctx_compress-s1-01342-twin
binledger
ctx_compress-s1-01342
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] restock_bin returns one item too few on every page. It lives in src/binledger/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 binledger layout. [tool_use kind=searc...
48,122
2,406
59
119
false
synth:8379501523015905309@2.33228
synth
[{"aliases": ["src/binledger/format.py", "format.py"], "kind": "file", "value": "src/binledger/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["src/binledger/cli.py", "cli.py"], "kind": "file", "value": "src/binledger/cli.py"}, {"aliase...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["data/stock_1.csv", "stock_1.csv"], "k...
ctx_compress-s1-01343
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] simplify_corridor raises KeyError for a missing key instead of returning the default. It lives in src/legworks/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 legwo...
34,862
1,743
72
87
false
synth:1727479418465318576@1
synth
[{"aliases": ["src/legworks/format.py", "format.py"], "kind": "file", "value": "src/legworks/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["src/legworks/audit.py", "audit.py"], "kind": "file", "value": "src/legworks/audit.py"}, {"alia...
[{"answers": ["src/legworks/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 mos...
ctx_compress-s1-01343-twin
legworks
ctx_compress-s1-01343
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_corridor raises KeyError for a missing key instead of returning the default. It lives in src/legworks/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 legwo...
17,655
1,000
40
87
false
synth:1727479418465318576@1
synth
[{"aliases": ["src/legworks/format.py", "format.py"], "kind": "file", "value": "src/legworks/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["src/legworks/audit.py", "audit.py"], "kind": "file", "value": "src/legworks/audit.py"}, {"alia...
[{"answers": ["src/legworks/format.py", "format.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's mos...
ctx_compress-s1-01344
indexflow
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] anchor_index rejects a range where lo equals hi, which is a valid single point range. It lives in src/indexflow/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 indexf...
97,478
3,072
125
140
false
synth:5607373180940063793@1.81943
synth
[{"aliases": ["src/indexflow/audit.py", "audit.py"], "kind": "file", "value": "src/indexflow/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["data/clause_8.csv", "clause_8.csv"], "kind": "file", "value": "data/clause_8.csv"}, {"aliases": ["...
[{"answers": ["src/indexflow/audit.py", "audit.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most...
ctx_compress-s1-01344-twin
indexflow
ctx_compress-s1-01344
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_index rejects a range where lo equals hi, which is a valid single point range. It lives in src/indexflow/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 indexf...
58,798
2,939
83
140
false
synth:5607373180940063793@1.81943
synth
[{"aliases": ["src/indexflow/audit.py", "audit.py"], "kind": "file", "value": "src/indexflow/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["data/clause_8.csv", "clause_8.csv"], "kind": "file", "value": "data/clause_8.csv"}, {"aliases": ["...
[{"answers": ["src/indexflow/audit.py", "audit.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most...
ctx_compress-s1-01345
beaconflow
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_beacon_all misses keys that differ only in case. It lives in src/beaconflow/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 beaconflow layout. [tool_use k...
66,135
3,072
74
93
false
synth:6391026693824377163@1.6098
synth
[{"aliases": ["src/beaconflow/store.py", "store.py"], "kind": "file", "value": "src/beaconflow/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/beaconflow/core.py", "core.py"], "kind": "file", "value": "src/beaconflow/core.py"}, {"alias...
[{"answers": ["src/beaconflow/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-s1-01345-twin
beaconflow
ctx_compress-s1-01345
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_beacon_all misses keys that differ only in case. It lives in src/beaconflow/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 beaconflow layout. [tool_use k...
43,231
2,161
48
93
false
synth:6391026693824377163@1.6098
synth
[{"aliases": ["src/beaconflow/store.py", "store.py"], "kind": "file", "value": "src/beaconflow/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/beaconflow/core.py", "core.py"], "kind": "file", "value": "src/beaconflow/core.py"}, {"alias...
[{"answers": ["src/beaconflow/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-s1-01346
ledgerdesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] settle_invoice returns one item too few on every page. It lives in src/ledgerdesk/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 ledgerdesk layout. [tool_use kind=...
68,963
3,072
89
110
false
synth:1158602678166064659@1.61708
synth
[{"aliases": ["src/ledgerdesk/compat.py", "compat.py"], "kind": "file", "value": "src/ledgerdesk/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["src/ledgerdesk/audit.py", "audit.py"], "kind": "file", "value": "src/ledgerdesk/audit.py"}...
[{"answers": ["src/ledgerdesk/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-s1-01346-twin
ledgerdesk
ctx_compress-s1-01346
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_invoice returns one item too few on every page. It lives in src/ledgerdesk/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 ledgerdesk layout. [tool_use kind=...
49,978
2,498
63
110
false
synth:1158602678166064659@1.61708
synth
[{"aliases": ["src/ledgerdesk/compat.py", "compat.py"], "kind": "file", "value": "src/ledgerdesk/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["src/ledgerdesk/audit.py", "audit.py"], "kind": "file", "value": "src/ledgerdesk/audit.py"}...
[{"answers": ["src/ledgerdesk/compat.py", "compat.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's m...
ctx_compress-s1-01347
sectionledger
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] paginate_heading ignores the cap, so late attempts wait for minutes. It lives in src/sectionledger/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 sectionledger layou...
99,668
3,072
97
120
false
synth:37148292165630345@2.36275
synth
[{"aliases": ["src/sectionledger/audit.py", "audit.py"], "kind": "file", "value": "src/sectionledger/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["src/sectionledger/io.py", "io.py"], "kind": "file", "value": "src/sectionledger/io.py"}, {...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["rules.py", "src/sectionledger/rules.p...
ctx_compress-s1-01347-twin
sectionledger
ctx_compress-s1-01347
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] paginate_heading ignores the cap, so late attempts wait for minutes. It lives in src/sectionledger/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 sectionledger layou...
56,908
2,845
69
120
false
synth:37148292165630345@2.36275
synth
[{"aliases": ["src/sectionledger/audit.py", "audit.py"], "kind": "file", "value": "src/sectionledger/audit.py"}, {"aliases": ["tests/test_audit.py", "test_audit.py"], "kind": "file", "value": "tests/test_audit.py"}, {"aliases": ["src/sectionledger/io.py", "io.py"], "kind": "file", "value": "src/sectionledger/io.py"}, {...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["data/draft_2.csv", "draft_2.csv"], "k...
ctx_compress-s1-01348
skuworks
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] age_stock raises ReorderDriftError for bytes input in strict mode instead of decoding it. It lives in src/skuworks/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 skuwork...
94,466
3,072
126
139
false
synth:8076358398886616340@1.76965
synth
[{"aliases": ["src/skuworks/cli.py", "cli.py"], "kind": "file", "value": "src/skuworks/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["data/pallet_12.csv", "pallet_12.csv"], "kind": "file", "value": "data/pallet_12.csv"}, {"aliases": ["src/skuwork...
[{"answers": ["src/skuworks/cli.py", "cli.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most rece...
ctx_compress-s1-01348-twin
skuworks
ctx_compress-s1-01348
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] age_stock raises ReorderDriftError for bytes input in strict mode instead of decoding it. It lives in src/skuworks/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 skuwork...
60,562
3,028
82
139
false
synth:8076358398886616340@1.76965
synth
[{"aliases": ["src/skuworks/cli.py", "cli.py"], "kind": "file", "value": "src/skuworks/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["data/pallet_12.csv", "pallet_12.csv"], "kind": "file", "value": "data/pallet_12.csv"}, {"aliases": ["src/skuwork...
[{"answers": ["src/skuworks/cli.py", "cli.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most rece...
ctx_compress-s1-01349
skuforge
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_shipment raises KeyError for a missing key instead of returning the default. It lives in src/skuforge/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 skuforge ...
49,977
2,498
80
91
false
synth:1675138697924583512@1.23745
synth
[{"aliases": ["src/skuforge/report.py", "report.py"], "kind": "file", "value": "src/skuforge/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["docs/changelog.md", "changelog.md"], "kind": "file", "value": "docs/CHANGELOG.md"}, {"aliases"...
[{"answers": ["src/skuforge/report.py", "report.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's mos...
ctx_compress-s1-01349-twin
skuforge
ctx_compress-s1-01349
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_shipment raises KeyError for a missing key instead of returning the default. It lives in src/skuforge/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 skuforge ...
20,380
1,019
41
91
false
synth:1675138697924583512@1.23745
synth
[{"aliases": ["src/skuforge/report.py", "report.py"], "kind": "file", "value": "src/skuforge/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["docs/changelog.md", "changelog.md"], "kind": "file", "value": "docs/CHANGELOG.md"}, {"aliases"...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["data/shipment_1.csv", "shipment_1.csv...
ctx_compress-s1-01350
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] normalise_glyph returns the upper bound for every value instead of clamping into the range. It lives in src/cuedesk/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 ...
84,800
3,072
115
142
false
synth:4612864938232204065@2.27793
synth
[{"aliases": ["src/cuedesk/format.py", "format.py"], "kind": "file", "value": "src/cuedesk/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["src/cuedesk/compat.py", "compat.py"], "kind": "file", "value": "src/cuedesk/compat.py"}, {"alias...
[{"answers": ["src/cuedesk/format.py", "format.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most...
ctx_compress-s1-01350-twin
cuedesk
ctx_compress-s1-01350
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_glyph returns the upper bound for every value instead of clamping into the range. It lives in src/cuedesk/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 ...
58,592
2,929
81
142
false
synth:4612864938232204065@2.27793
synth
[{"aliases": ["src/cuedesk/format.py", "format.py"], "kind": "file", "value": "src/cuedesk/format.py"}, {"aliases": ["tests/test_format.py", "test_format.py"], "kind": "file", "value": "tests/test_format.py"}, {"aliases": ["src/cuedesk/compat.py", "compat.py"], "kind": "file", "value": "src/cuedesk/compat.py"}, {"alias...
[{"answers": ["src/cuedesk/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 most...
ctx_compress-s1-01351
creditforge
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] accrue_credit_window produces doubled dashes for consecutive separators. It lives in src/creditforge/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 creditforge layou...
54,461
2,723
95
106
false
synth:2219126171311131269@1.41253
synth
[{"aliases": ["src/creditforge/store.py", "store.py"], "kind": "file", "value": "src/creditforge/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/creditforge/core.py", "core.py"], "kind": "file", "value": "src/creditforge/core.py"}, {"a...
[{"answers": ["src/creditforge/store.py", "store.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's mo...
ctx_compress-s1-01351-twin
creditforge
ctx_compress-s1-01351
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] accrue_credit_window produces doubled dashes for consecutive separators. It lives in src/creditforge/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 creditforge layou...
33,007
1,650
61
106
false
synth:2219126171311131269@1.41253
synth
[{"aliases": ["src/creditforge/store.py", "store.py"], "kind": "file", "value": "src/creditforge/store.py"}, {"aliases": ["tests/test_store.py", "test_store.py"], "kind": "file", "value": "tests/test_store.py"}, {"aliases": ["src/creditforge/core.py", "core.py"], "kind": "file", "value": "src/creditforge/core.py"}, {"a...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["io.py", "src/creditforge/io.py"], "ki...
ctx_compress-s1-01352
snippetdesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] tokenise_stem raises KeyError for a missing key instead of returning the default. It lives in src/snippetdesk/core.py; `python -m pytest tests/test_core.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the snippetdes...
100,732
3,072
96
115
false
synth:3468140042366540314@2.0182
synth
[{"aliases": ["src/snippetdesk/core.py", "core.py"], "kind": "file", "value": "src/snippetdesk/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["data/cue_3.csv", "cue_3.csv"], "kind": "file", "value": "data/cue_3.csv"}, {"aliases": ["data/lexeme...
[{"answers": ["src/snippetdesk/core.py", "core.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most...
ctx_compress-s1-01352-twin
snippetdesk
ctx_compress-s1-01352
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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 raises KeyError for a missing key instead of returning the default. It lives in src/snippetdesk/core.py; `python -m pytest tests/test_core.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping the snippetdes...
59,093
2,954
68
115
false
synth:3468140042366540314@2.0182
synth
[{"aliases": ["src/snippetdesk/core.py", "core.py"], "kind": "file", "value": "src/snippetdesk/core.py"}, {"aliases": ["tests/test_core.py", "test_core.py"], "kind": "file", "value": "tests/test_core.py"}, {"aliases": ["data/cue_3.csv", "cue_3.csv"], "kind": "file", "value": "data/cue_3.csv"}, {"aliases": ["data/lexeme...
[{"answers": ["src/snippetdesk/core.py", "core.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["fail"], "question": "Did the agent's most...
ctx_compress-s1-01353
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_beacon raises ZeroDivisionError on the first value and lags by one afterwards. It lives in src/waypointworks/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 waypoint...
44,861
2,243
88
97
false
synth:8991234655119055466@1.09357
synth
[{"aliases": ["src/waypointworks/cli.py", "cli.py"], "kind": "file", "value": "src/waypointworks/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["src/waypointworks/schema.py", "schema.py"], "kind": "file", "value": "src/waypointworks/schema.py"}, {...
[{"answers": ["src/waypointworks/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-s1-01353-twin
waypointworks
ctx_compress-s1-01353
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_beacon raises ZeroDivisionError on the first value and lags by one afterwards. It lives in src/waypointworks/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 waypoint...
27,663
1,383
52
97
false
synth:8991234655119055466@1.09357
synth
[{"aliases": ["src/waypointworks/cli.py", "cli.py"], "kind": "file", "value": "src/waypointworks/cli.py"}, {"aliases": ["tests/test_cli.py", "test_cli.py"], "kind": "file", "value": "tests/test_cli.py"}, {"aliases": ["src/waypointworks/schema.py", "schema.py"], "kind": "file", "value": "src/waypointworks/schema.py"}, {...
[{"answers": ["src/waypointworks/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-s1-01354
windowkit
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_slot_strict rejects a range where lo equals hi, which is a valid single point range. It lives in src/windowkit/compat.py; `python -m pytest tests/test_compat.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping t...
83,165
3,072
107
122
false
synth:3809997107633702183@1.52585
synth
[{"aliases": ["src/windowkit/compat.py", "compat.py"], "kind": "file", "value": "src/windowkit/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["src/windowkit/report.py", "report.py"], "kind": "file", "value": "src/windowkit/report.py"},...
[{"answers": ["src/windowkit/compat.py", "compat.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's mo...
ctx_compress-s1-01354-twin
windowkit
ctx_compress-s1-01354
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_slot_strict rejects a range where lo equals hi, which is a valid single point range. It lives in src/windowkit/compat.py; `python -m pytest tests/test_compat.py -q` shows the failures. Fix it without changing the public signature and add a regression test. [assistant] I will start by mapping t...
47,132
2,356
69
122
false
synth:3809997107633702183@1.52585
synth
[{"aliases": ["src/windowkit/compat.py", "compat.py"], "kind": "file", "value": "src/windowkit/compat.py"}, {"aliases": ["tests/test_compat.py", "test_compat.py"], "kind": "file", "value": "tests/test_compat.py"}, {"aliases": ["src/windowkit/report.py", "report.py"], "kind": "file", "value": "src/windowkit/report.py"},...
[{"answers": ["src/windowkit/compat.py", "compat.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's mo...
ctx_compress-s1-01355
bookingledger
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_slot_batch averages one value too many once the window is full. It lives in src/bookingledger/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 bookingledger la...
41,835
2,091
73
92
false
synth:1073351167600267536@1.4164
synth
[{"aliases": ["src/bookingledger/report.py", "report.py"], "kind": "file", "value": "src/bookingledger/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["src/bookingledger/schema.py", "schema.py"], "kind": "file", "value": "src/bookingled...
[{"answers": ["src/bookingledger/report.py", "report.py"], "question": "Which file did the agent edit most recently?", "type": "exact"}, {"answers": ["yes"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent'...
ctx_compress-s1-01355-twin
bookingledger
ctx_compress-s1-01355
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their 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_slot_batch averages one value too many once the window is full. It lives in src/bookingledger/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 bookingledger la...
23,213
1,160
43
92
false
synth:1073351167600267536@1.4164
synth
[{"aliases": ["src/bookingledger/report.py", "report.py"], "kind": "file", "value": "src/bookingledger/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["src/bookingledger/schema.py", "schema.py"], "kind": "file", "value": "src/bookingled...
[{"answers": ["no"], "question": "Has the agent changed any existing file in the repository yet? Answer yes or no.", "type": "word"}, {"answers": ["pass"], "question": "Did the agent's most recent test run pass or fail? Answer pass or fail.", "type": "word"}, {"answers": {"file": ["audit.py", "src/bookingledger/audit.p...
ctx_compress-s1-01356
counterdesk
You compact a coding agent's session so the agent can continue from your summary alone. The session is inside <session> tags: [user] messages, [assistant] text, [tool_use] calls with their kind and file, and [tool_result] outputs. Write the summary in plain markdown with exactly these nine headings, in this order, each...
<session> [user] merge_series raises KeyError for a missing key instead of returning the default. It lives in src/counterdesk/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 counter...
65,138
3,072
103
124
false
synth:3243211340495274521@2.37398
synth
[{"aliases": ["src/counterdesk/report.py", "report.py"], "kind": "file", "value": "src/counterdesk/report.py"}, {"aliases": ["tests/test_report.py", "test_report.py"], "kind": "file", "value": "tests/test_report.py"}, {"aliases": ["src/counterdesk/limits.py", "limits.py"], "kind": "file", "value": "src/counterdesk/limi...
[{"answers": ["src/counterdesk/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 ...