89975a1fd6
fix(trekendsession): Phase 3 re-reads the arguments instead of splicing the label into the shell line
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-26 16:12:44 +01:00
2c0a98c1ba
fix(commands): Phase 0 gives the arguments to the parser on stdin, never as shell words
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-26 15:55:21 +01:00
f1704f0800
fix(arg-parser): take the arguments as stdin text, and block a heredoc that closes early
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-26 15:40:21 +01:00
738efa3e24
fix(review): the coordinator and orchestrator texts say what the code does on refutation and the brief-conflict exception
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-26 15:18:03 +01:00
878871a6ad
fix(review): the Arm B workflow keeps the contract's row on a trimmed title and a tie
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-26 15:06:23 +01:00
9145d2bbf5
fix(trekexecute): the back-to-planning save line reads the whole diff from a subdirectory
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-26 14:53:26 +01:00
afe56899a9
fix(park): park from a CRLF diff and say that a staged new file is parked
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-26 14:44:33 +01:00
bf4736a254
fix(park): read HEAD membership once and move nothing when it cannot be read
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-26 14:36:03 +01:00
5ddbcd3de6
fix(verdict-gate): free text added inside the lessons block makes the review stale
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-26 14:11:41 +01:00
77544841d7
fix(verdict-gate): a failed git call is reported as not measured, not stale
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-26 14:00:42 +01:00
703c508433
fix(verdict-gate): read --repo=<dir> and refuse unknown options
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-26 13:44:11 +01:00
e279f12131
fix(verification): the done line and the release check take only full commit ids
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-26 13:33:51 +01:00
cf7af30b46
fix(verdict-gate): accept only a full 40- or 64-hex reviewed commit
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-26 13:13:22 +01:00
7566b71674
refactor(red-proof): the 4.5a script is a wrapper over the shared one; add the 4.5b wrapper
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-26 12:58:23 +01:00
2dbfbaf8ca
feat(red-proof): one red script that takes base, ids, mutations and files as arguments
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-26 12:38:44 +01:00
cde4c1d3ed
fix(red-proof): the judge skips other rounds' ids and reports an unreadable tag
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-26 12:19:36 +01:00
be0b631fed
merge: fix round 4.5a — the serious findings, decision 67 and the too-lax 6.5.0 Known issues (decisions 57, 67, 68, 78)
...
Review WARN (0 BLOCKER, 13 MAJOR, 4 MINOR), verdict gate exit 0; the 13 MAJOR go first in hardening part 4.5b (operator 2026-10-25).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-25 16:21:02 +01:00
af90e82794
refactor: simplify pass (trekexecute)
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-24 22:22:18 +02:00
f6bdc2de59
fix(docs): the put-back, the research triggers and Pass 3's high-effort skip say what the code does
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-24 21:19:58 +02:00
8a2a989f79
fix(siblings): the texts found next to round 4.5a's fixes say what the code does
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-24 20:41:53 +02:00
e351e7c29d
docs(changelog): fix round 4.5a
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-24 20:11:22 +02:00
5476a36782
test(red): prove every id of round 4.5a red at ab6b32c, with a judge that is tested to go red
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-24 19:23:59 +02:00
7869c4cbae
fix(execute): the park step parks intent-to-add and coloured-diff files from any directory and exits 2 on a git failure
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 21:00:24 +02:00
773aa44dc9
fix(review): the workflow floor and put-back match each finding to its own row
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 19:59:17 +02:00
a5fcb5e8b4
fix(review): group a triplet's findings title first, so arrival order never changes the rows
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 18:30:52 +02:00
7175783ce5
fix(review): a past-the-end line makes a finding unverified, never refuted, and the Declined to Judge copy-through is pinned
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 17:13:03 +02:00
761f406faa
fix(review): changed code and instruction files get deep review by default, summary-only is for docs
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 15:37:39 +02:00
b16e39ccd9
fix(research): the contrarian runs as a second wave with the leading answer to attack
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 15:23:18 +02:00
f32baf50ad
fix(continue): run the arg-parser, find the master prompt at the repo root, and let the consistency check read the state file
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 15:05:19 +02:00
403afe606f
fix(endsession): parse arguments with the arg-parser so a slug with -h is no help request, and only the operator starts it
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 14:45:25 +02:00
6226372eaa
fix(execute): the final report says whether the app was seen with Claude Code's run
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 14:26:28 +02:00
050a77f513
fix(execute): the merge lock refuses a review of an older commit (review-stale)
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 11:06:14 +02:00
6c0718c8d8
fix(review): the verdict gate refuses, with --repo, a review of another commit than the one being merged
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-23 10:48:29 +02:00
a9f5f6251a
fix(execute): the completion gate says which checks are commands and which are executor judgement
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-22 16:25:23 +02:00
2fe6feb74d
fix(execute): run the brief check in Phase 1, and say what effort really changes
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-22 16:07:27 +02:00
2b5d5f72bc
fix(execute): name and read --profile and --gates in the hint, the usage block and Phase 1
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-22 15:50:13 +02:00
af36578f12
fix(gates): one --gates rule in every command, run through the shim, and no executor pause the executor lacks
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-22 15:38:20 +02:00
e2fcf4060c
fix(gates): the shim reads a bare --gates and closed as on, and never swallows the next flag
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-22 15:20:17 +02:00
8f9b0abfad
fix(brief): a brief with no research topics prints its next steps instead of entering the auto-only phase
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-22 15:12:59 +02:00
95eaa8dc57
fix(stats): no command promises a stats key its record does not write
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-22 15:04:29 +02:00
074feef475
fix(plan): every planning and research spawn takes the resolved model, and no turn budget is promised
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-22 14:51:04 +02:00
eafe023c74
fix(plan): execute hands the plan to /trekexecute instead of building it in the planning session
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-22 14:34:57 +02:00
be83e6e066
fix(docs): name /trekplan where the text sent readers to a reference document, and give the reference documents read-only tools
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-22 14:23:05 +02:00
4312882451
fix(agents): reviewers drop a finding only with a Declined to Judge line, research-scout gets Tavily, no dead override reference
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-22 14:11:47 +02:00
ab6b32c930
merge: wave 4 order B — voyage:annotating (decision 61, 62)
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-19 20:41:46 +02:00
45eaedc6b4
merge: main (wave 4 order A) into order B before B goes into main
...
Conflicts on the union lines only: CHANGELOG Unreleased (both orders'
entries kept), CLAUDE.md (/trekexecute row from A, /trekreview row from
B), tests/fixtures/jsonl-schemas.md (trekplan line_ref from B,
trekexecute line_ref from A). Suite on the merged tree 2682 (2680/0/2).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-19 18:49:02 +02:00
074b44a7e0
merge: wave 4 order A — P10's model-free part and the run coupling (decision 61, 63, 64)
...
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-19 18:44:29 +02:00
843897ae33
feat(skills): voyage:annotating, the operator-annotation loop as one skill (order B)
...
The three producing commands name voyage:annotating where they build the page; the description of the page and the freehand revision left them. The skill: the person writes every note, Fix/Change edit only the quoted text, a Question is answered and changes nothing, the approval an edit costs is said before the edit, the page is rebuilt. Research documents on request (decision 62). Red baseline 3 runs per scenario without and with the skill; six eval cases; 11 pins.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-19 18:36:00 +02:00
1989656a9c
feat(run): the run step names Claude Code's run; owner=run is a built-in owner (A2)
...
The coupling to Claude Code's built-in run skill (wave 4, order A2,
decisions 54 and 64). The plan's "run the thing" step, the executor's
End to end part and the plan template each say in one sentence that the
app is seen and driven with run, and that the verdict is still the
End-to-end: command, because run gives no exit code. The template quotes
the name instead of backticking it: every backticked span on a template
step bullet is read as a runnable example.
R2: lib/skills/routing.mjs keeps a short fixed list of Claude Code's own
skills a quiet case may name as owner (BUILTIN_OWNERS, only run, seen in
2.1.303). For such an owner only "not taken" is checked, and the report
says so; ROUTING_NEGATIVE_WON stands, and an owner on no list is still
ROUTING_UNKNOWN_OWNER. skill-writing-criteria-quiet-2 gets owner=run.
The trekplan.md paragraph is reflowed to its old line count so the stats
line_ref there does not move (order B edits the same file); the
trekexecute Phase 9 line_ref fixture moves by two lines.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-19 18:26:10 +02:00
cf08df9d55
feat(routing): every eval case names its event; the P10 eval gate is exempted with its reason (A1)
...
P10's model-free part (wave 4, order A1, decision 63). The routing test now
checks every case in evals/ for from=, not only the skill-tagged ones (R1):
a case with no skill tag is not routed, only checked, and the report gives
both counts ("45 cases checked for from=, 37 routed"). The eight test-set
cases get from=proevesett-eval-suite (d311f3b3892a83
2026-10-19 18:15:00 +02:00
