Skip to content

docs(skills): rename prepare-pr, slim it and match current ci - #16740

Merged
chenmingwei23 merged 1 commit into
mainfrom
skills/prepare-pr-slim
Oct 4, 2026
Merged

chenmingwei23 merged 1 commit into
mainfrom
skills/prepare-pr-slim

Conversation

@bolichen97

@bolichen97 bolichen97 commented Oct 4, 2026 •

Copy link
Copy Markdown
Contributor

Problem / Motivation

Goal: Every Kiro Crew author's agent ships a PR with a prepare-pr skill that matches today's CI and states the repository's process, in a file short enough to read on every load.

The skill had drifted from the CI it drives. It never named Security Scope Review or Internal Content Scan, listed one of seven override targets, missed PR Hygiene's ## Pattern harvest and Reader: rules, and still sent deferrals through needs-triage, which Issue Gate stopped reading in #16697. It told an outside author nothing about issues, triage or tiers. It sat at its 68000-byte budget, and its description did not say "Kiro Crew PR", so the agent did not always load it.

Why it matters

Each missing rule costs an author a CI round to learn, and an outside contributor cannot learn the issue and tier process from the loop at all. Main stays red until someone happens to fix it, and green PRs decay unmerged.

Not a goal

  • Using it on another repository or on a CRUX code review.
  • Changing any script, workflow or CI rule; this PR only describes them.
  • Turning on Issue Gate or the merge queue.

What changed (motivation → approach → change)

The skill now carries the rules an author needs, and the loop finishes the job instead of stopping at review-ready.

Area Now
Name and scope renamed to kirocrew-prepare-pr (directory, frontmatter, and every reference in CI, docs, skills and tests); repo_scope: src/kiro_crew loads it in every Kiro Crew session and hides it everywhere else; the CRUX cr triggers are gone. An install that holds the old kirocrew-dev/prepare-pr copy has it quarantined on upgrade, and skill_runtime/search.py resolves a saved $prepare-pr (a hook, an AGENTS.md line) to the new name
CI at a glance issue → Captain triage → tier:T1..tier:T4 and which tiers wait for triaged; every blocking lane and what to do when it is red; PR Hygiene; the merge queue
Red main a red that also fails on main gets a search for an existing fix, else a tier:T1 issue and its own fix PR; the original PR waits, then rebases
Landing the full loop arms auto-merge by default and keeps watching up to 24 hours, rebasing conflicts (max_cycles=280)
Overrides a writer's agent may override a false positive or over-engineering finding itself, with a reason starting agent:, and must report it; security, data-loss, corruption, crash and removed-guard findings always go to a person. docs/ci/ci-and-reviews.md § Human override says the same

The rename rewrites only the name wherever it appears: .github/ workflows and scripts, CONTRIBUTING.md, .gitignore comments, src/kiro_crew/agent.py, src/kiro_crew/frontmatter.py, src/kiro_crew/security_posture.py, src/kiro_crew/docs/agents.md, other skills, specs and tests; src/kiro_crew/skills.py maps both old names to the new key.

Repeated text moved out to --help, the PR template, the babysit skill and references/rationale.md, which also records why each new rule exists. SKILL.md drops from 67911 to about 55000 bytes and its ceiling test ratchets to 56000. docs/ci/ci-and-reviews.md now names the tier-based deferral path. RFCs keep the old name: they are history.

Backwards compatibility

Breaking: an agent spec that maps this skill by path, written by the crew page as skill://~/.kiro/crew/skills/kirocrew-dev/prepare-pr/SKILL.md, loads nothing after upgrade, because the old SKILL.md is quarantined so two copies never load. Re-map the skill from the crew page to fix it. Writer sweep on origin/main @ 5bb27bd: no shipped agent template or spec maps it by path; src/kiro_crew/dashboard/handlers/_shared.py:_skill_uri_for_path writes the URI only when an operator maps a skill. A typed or hook-saved $prepare-pr keeps working through the alias, and pr-readiness.yml and Issue Gate read the evaluator from the default branch, so each side always finds its own path.

Reader: src/kiro_crew/agent_discovery.py:expand_skill_uri -- chat, subagent -- a saved old-path URI resolves to the quarantined file; the operator re-maps the skill
Reader: src/kiro_crew/skill_runtime/search.py:resolve_dollar_skills -- chat, cron -- $prepare-pr resolves through _RELOCATED_SKILLS; test_renamed_builtin_answers_to_its_old_name
Reader: test/test_review_repair_routing_skill.py:test_comment_aware_recipes_are_finite_ungated_and_keep_repair_routing -- chat -- pins the new (280, 86400) recipe
Reader: test/test_pr_quality_gates.py:TestScreenshotEvidence -- chat -- the pinned screenshot bullets are kept verbatim

Tests

test_renamed_nested_copy_quarantined_when_new_name_present pins the upgrade path for the old nested copy; test_renamed_builtin_answers_to_its_old_name and test_installed_skill_keeps_its_name_over_a_relocation_alias pin the $prepare-pr alias and that a user's own skill of that name still wins. Otherwise no new tests: every rule the skill text carries is already pinned by about 150 assertions across 18 test files, and all of them pass on this head. The ceiling and recipe pins are updated to the new values, and so are the security-conductor spec digests in test_agent_refactor_spec_bytes.py / test_agent_refactor_contracts.py, whose prompt names the skill.

Manual verification

  • Ran every test file that reads prepare-pr/SKILL.md or its references: green.
  • Ran scripts/docs-lint.sh: green.
  • Local review by the profile's gpt and opus reviewers; the one Opus blocker (agent self-override on security findings) is fixed and re-verified.

Related Issues

Closes #16731

Checklist

  • At most two commits (one is the norm), with a Conventional Commits title (feat|fix|docs|style|refactor|perf|test|chore|ci|build|revert: ...)
  • Existing tests pass and new tests added for new functionality
  • Self-review completed; code follows project style guidelines
  • Documentation updated (if applicable)
  • No secrets, credentials, or internal references in the diff

@bolichen97
bolichen97 requested a review from a team as a code owner October 4, 2026 09:35
@bolichen97
bolichen97 requested a review from iamwhatever October 4, 2026 09:35
@github-actions github-actions Bot added the readiness: checking Automated validation is still running label Oct 4, 2026
@github-actions

github-actions Bot commented Oct 4, 2026 •

Copy link
Copy Markdown
Contributor

Design Review (Opus 5.5) — 🟡 CONCERNS

Design-level review of 608b4aa8acba5006ca8d5ac18fd867ee95ddcbc7 — updated in place on each push. A BLOCK verdict blocks PR readiness; PASS/CONCERNS are advisory.

Design-Verdict: CONCERNS

An agent can now override AI-review findings on its own PR, and the loop arms auto-merge by default. That adds a policy to a PR whose goal is a rename and a "describe-only" update.

Watch

  • The agent can now clear the gate on its own PR. The skill now says "may override … without asking" and arms auto-merge by default. The docs change "proof a human adjudicated it" to "proof a writer adjudicated it (… through their agent)". The exclusions (security, data-loss, crash, removed-guard) and the agent: prefix are written guidance only. pr-readiness checks only that the commenter has write access, and the agent decides which category its own finding falls in. So a writer's agent can rebut a design or GPT BLOCK, override it and auto-merge, and no person sees the finding. That contradicts "Not a goal: Changing any … CI rule; this PR only describes them" and "Human judgment is the final authority". It can be reverted, but it changes the review process.
    Clears when: the agent-override rule and the auto-merge default move to their own PR, with a maintainer-written docs/decisions entry; or a maintainer approves them on this PR.
  • Operators who mapped the old skill by path lose it with no notice. The description accepts that a saved skill://…/kirocrew-dev/prepare-pr/SKILL.md "loads nothing after upgrade". The quarantine only logs this on the server side, so nothing tells the operator the skill is gone.
    Clears when: the quarantine rewrites a matching old URI in agent specs to the new path, or the crew page flags the dead mapping.

[DESIGN-REVIEWED] 608b4aa

Description read

[DESCRIPTION-READ] 6aa8164ea868be0b55da2e0a3997366d481408e6cdd1488da2327fe3b4b800da

That sha256 names everything this verdict read: the PR description, plus the 1 evidence file(s) the workflow collected from it and handed this review. It covers that evidence because the media strip replaces every attachment URL with the same placeholder, so an edit swapping one attachment for another would leave the description's captured text identical while what this review judged changed. Recompute it with .github/scripts/pr-description-capture.sh over the same inputs; if it differs, the description or that evidence changed after this verdict was formed and any finding drawn from either is unproven.

@github-actions github-actions Bot added readiness: action required A blocking check or review needs attention and removed readiness: checking Automated validation is still running labels Oct 4, 2026
@github-actions

github-actions Bot commented Oct 4, 2026 •

Copy link
Copy Markdown
Contributor

First Principles Review (Opus 5.5) — ✅ human override accepted

@bolichen97 overrode this lane for 608b4aa8acba5006ca8d5ac18fd867ee95ddcbc7 via /ai-review override (this commit only).

@github-actions

github-actions Bot commented Oct 4, 2026 •

Copy link
Copy Markdown
Contributor

Opus 5.5 Review — ✅ no blocking findings

Reviewed 608b4aa8acba5006ca8d5ac18fd867ee95ddcbc7 — this comment is updated in place on each push.

Review details

No findings block this merge. One advisory finding below.

FINDING — src/kiro_crew/builtin_skills/kirocrew-dev/kirocrew-prepare-pr/SKILL.md:24 — The trigger list now includes "fix ci", "make it green" and "commit and push". With these, a bare "fix ci" falls through to the default "Full loop", and the Guardrails at line 293 say "A request to run this loop authorizes its own pushes, … auto-merge, and the red-main fix PR it opens". So the agent pushes, arms auto-merge and opens new PRs even though nobody asked it to push. That contradicts AGENTS.md ("Do NOT git push unless the user explicitly says to push") and the same PR's kirocrew-worktree-dev ("A fix-and-push monitoring scope must be explicit; absent it, confirm each push"). It also drops the base rule Generic remediation ("fix CI", "make it green") is not a ship request. → Fix: restore that sentence, and limit the Guardrails' self-authorization to explicit ship or land requests.

[OPUS-REVIEWED] 608b4aa

Verdict parsed from the review's SHA-scoped output markers for commit 608b4aa8acba5006ca8d5ac18fd867ee95ddcbc7.

False positive or not applicable? A repository writer can comment:
/ai-review override fable 608b4aa8acba5006ca8d5ac18fd867ee95ddcbc7: <one-sentence reason>

@bolichen97
bolichen97 force-pushed the skills/prepare-pr-slim branch from 374b615 to 8214759 Compare October 4, 2026 09:44
@github-actions

github-actions Bot commented Oct 4, 2026 •

Copy link
Copy Markdown
Contributor

GPT 6.1 Review — ✅ no blocking findings

GPT 6.1 completed its review of 608b4aa8acba5006ca8d5ac18fd867ee95ddcbc7 and found no blocking issues.

This comment is updated in place on each push.

Review details

No findings.
[GPT-REVIEWED] 608b4aa

False positive or not applicable? A repository writer can comment:
/ai-review override gpt 608b4aa8acba5006ca8d5ac18fd867ee95ddcbc7: <one-sentence reason>

@github-actions github-actions Bot added readiness: checking Automated validation is still running and removed readiness: action required A blocking check or review needs attention labels Oct 4, 2026
@bolichen97 bolichen97 changed the title docs(prepare-pr): slim the skill and bring it up to current ci docs(skills): rename prepare-pr, slim it and match current ci Oct 4, 2026
@bolichen97
bolichen97 force-pushed the skills/prepare-pr-slim branch from 8214759 to 3828984 Compare October 4, 2026 10:10
@bolichen97

Copy link
Copy Markdown
Contributor Author

Frozen-goal edit, at the maintainer's request. @bolichen97 asked for the skill to be renamed to kirocrew-prepare-pr, with every template, hook, CI and doc reference updated. The first ## Not a goal bullet ("Renaming the skill or its directory") contradicted that, so it was removed. The **Goal:** line and ## Why it matters are unchanged.

The rename keeps working for existing setups: _RELOCATED_SKILLS quarantines an installed kirocrew-dev/prepare-pr copy on upgrade, and a saved $prepare-pr still resolves to the new skill unless a user skill already owns that name.

@github-actions

github-actions Bot commented Oct 4, 2026 •

Copy link
Copy Markdown
Contributor

Security Scope Review -- 🟡 CONCERNS

Which legitimate operations does 608b4aa8acba5006ca8d5ac18fd867ee95ddcbc7 newly refuse? Updated in place on each push. A script-confirmed regression blocks; the model proposes candidates and never decides a refusal.

Verdict basis: the reviewer's own judgement; no confirmed regression.

Security scope review -- adjudicated candidates

  • macos-latest (platform posix): 12 classified, 0 newly refused, 12 still allowed.
  • ubuntu-latest (platform posix): 12 classified, 0 newly refused, 12 still allowed.
  • windows-latest (platform windows): 6 classified, 0 newly refused, 6 still allowed.

No candidate flipped from allowed to refused.

Scope-Verdict: CONCERNS

TIGHTENING: built-in skill relocation/quarantine in _ensure_builtin_skills (src/kiro_crew/skills.py) -- it refuses to rename a stale skill copy through a linked directory, so a file the operator linked in from outside the skills home is never renamed: + if is_link_or_junction(base.joinpath(*Path(old_name).parts[: i + 1])) ... logger.warning("Skill %s relocated to %s, but %s is a link; not quarantining through it (the linked copy is untouched)" ...); continue. No hunk touches the tool-call deny composite (path fence, sensitive tier, exfil shapes or rule catalog). For shell commands, the only changes are path renames: kirocrew-dev/prepare-pr/ becomes kirocrew-dev/kirocrew-prepare-pr/, including the NON_EGRESS_REDACTION_MODULES entry and the four push_guard corpus rows.

ADJUDICATED: 18 candidates, verdicts pending from the harness (none were handed to me in this run, so I report no confirmed newly-refused rows)

PLATFORM COVERAGE:

  • posix (11 rows): every script the renamed SKILL.md runs through SKILL_DIR="$HOME/.kiro/crew/skills/kirocrew-dev/kirocrew-prepare-pr":
    • push_guard --commit, --squash and --require-single-on-base
    • resolve_profile, preflight, pr_status, pr_findings, green_age and local_review
    • the two source-tree forms named in the diff: prove.py and the pr-readiness GATE path
    • a read-only listing of the new install path
  • windows (7 rows):
    • %USERPROFILE%\… with python
    • the py -3 launcher
    • .venv\Scripts\python.exe against a backslash source path
    • an absolute C:\Users\… drive-letter path
    • a PowerShell $env:USERPROFILE listing
  • any (0 rows): every row carries a path, so none is platform-neutral.
  • Not probed: a UNC-path spelling of the skills home. The diff adds no UNC-specific assumption.

UNADJUDICATED:

  • Guard: the link/junction screen in _ensure_builtin_skills (skills.py). It is a filesystem check at gateway startup, not a shell command line, so the classifier cannot give it a verdict.

  • Call path: gateway startup → _ensure_builtin_skills(base) → loop over _RELOCATED_SKILLS → for each old_name whose base/old_name/SKILL.md exists and whose new SKILL.md exists → is_link_or_junction on each prefix of old_name → skip the quarantine with a warning.

  • Legitimate inputs this newly leaves unmigrated (they would have been my candidates):

    1. An operator who symlinked ~/.kiro/crew/skills/kirocrew-dev to a shared dotfiles checkout, which is a common setup. The new kirocrew-dev/prepare-pr → kirocrew-dev/kirocrew-prepare-pr entry now never quarantines the old nested copy, so both kirocrew-dev/prepare-pr and kirocrew-dev/kirocrew-prepare-pr stay loaded.
    2. The same operator on Windows with a directory junction at %USERPROFILE%\.kiro\crew\skills\kirocrew-dev.
    3. A flat prepare-pr that is itself a link. This was arguably refused correctly before.

    In cases 1 and 2 the old copy stays loadable, and it carries stale script paths that no longer exist in the packaged tree. Whether the loader then prefers the stale copy is something this lane cannot settle.

  • Second guard: the new $prepare-pr fallback in skill_runtime/search.py resolve_dollar_skills. It loosens resolution rather than tightening it, but it has a refusal edge: not leaves.get(token.casefold()) leaves an old name unresolved whenever any installed leaf holds it. In case 1, the un-quarantined kirocrew-dev/prepare-pr leaf still exists, so the fallback never fires there. That input path is an agent prompt or hook skills[] value, not a shell command, so it has no verdict either.

[SCOPE-REVIEWED] 608b4aa

@bolichen97
bolichen97 force-pushed the skills/prepare-pr-slim branch 2 times, most recently from ab1cc58 to 3088197 Compare October 4, 2026 10:34
@bolichen97

Copy link
Copy Markdown
Contributor Author

/ai-review override first-principles 3088197: agent: records maintainer @bolichen97's explicit decisions for #16731 - full loop arms auto-merge by default with a 24h watch, a writer's agent may override false-positive/over-engineering findings (never security-class) and must report it, red main gets its own T1 fix PR, the skill is renamed kirocrew-prepare-pr and scoped to this repo.

@github-actions

github-actions Bot commented Oct 4, 2026

Copy link
Copy Markdown
Contributor

Human judgment recorded

@bolichen97 marked the first-principles AI finding as false positive, not applicable, or explicitly accepted for 3088197dd688b0eb1d85efd0d9a009a6cc8dc03e.

agent: records maintainer @bolichen97's explicit decisions for #16731 - full loop arms auto-merge by default with a 24h watch, a writer's agent may override false-positive/over-engineering findings (never security-class) and must report it, red main gets its own T1 fix PR, the skill is renamed kirocrew-prepare-pr and scoped to this repo.

This decision applies only to this commit. A new push requires a new judgment.

@bolichen97
bolichen97 force-pushed the skills/prepare-pr-slim branch from 3088197 to 8ba217e Compare October 4, 2026 10:50
@bolichen97

Copy link
Copy Markdown
Contributor Author

/ai-review override first-principles 8ba217e: agent: records maintainer @bolichen97's explicit decisions for #16731 - full loop arms auto-merge by default with a 24h watch, a writer's agent may override false-positive/over-engineering findings (never security-class) and must report it, red main gets its own T1 fix PR, the skill is renamed kirocrew-prepare-pr and scoped to this repo.

@github-actions

github-actions Bot commented Oct 4, 2026

Copy link
Copy Markdown
Contributor

Human judgment recorded

@bolichen97 marked the first-principles AI finding as false positive, not applicable, or explicitly accepted for 8ba217edeab2239b31b5d25dd5440e5c658ce025.

agent: records maintainer @bolichen97's explicit decisions for #16731 - full loop arms auto-merge by default with a 24h watch, a writer's agent may override false-positive/over-engineering findings (never security-class) and must report it, red main gets its own T1 fix PR, the skill is renamed kirocrew-prepare-pr and scoped to this repo.

This decision applies only to this commit. A new push requires a new judgment.

@github-actions

github-actions Bot commented Oct 4, 2026

Copy link
Copy Markdown
Contributor

Human judgment recorded, but the following reviewer lane(s) could not be re-run automatically: First Principles Review. Re-run the lane's latest workflow run for 8ba217edeab2239b31b5d25dd5440e5c658ce025 manually from the Actions tab.

@bolichen97

bolichen97 commented Oct 4, 2026 •

Copy link
Copy Markdown
Contributor Author
  • span=c2864a047df5 needs-a-decision: agent-override policy

The policy is the maintainer's own request (@bolichen97, #16731): a writer's agent may override false-positive and over-engineering findings and must report each one.
The carve-out is in the skill and in ci-and-reviews.md: never a security, data-loss, corruption, crash or removed-guard finding; those go to a person.
The record stays honest: the reason starts with agent:, the writer whose account posts it is accountable, and the authorization section says so.
Putting it into docs/decisions/ needs the maintainer to restate it on this PR (decisions rule 4); that is his call, asked in the session.

@bolichen97

bolichen97 commented Oct 4, 2026 •

Copy link
Copy Markdown
Contributor Author
  • span=a5479b0d8b4d rebutted: split the rename from the policy changes

Splitting buys nothing the maintainer has not already ruled on: both policies are his explicit decisions for this PR (#16731), recorded by the First Principles override on this head.
A split costs two more full CI and review rounds for the same diff; the rename alone already needed four of them.

@bolichen97

bolichen97 commented Oct 4, 2026 •

Copy link
Copy Markdown
Contributor Author
  • span=6d18f7e9f97d fixed: ambiguous leaf no longer falls through to the relocation alias

Real: two installed skills sharing a leaf left matched None, and the alias then injected the built-in.
Fixed in 316dc0e830: the alias applies only when the token is neither an installed key nor any installed leaf, so an ambiguous leaf stays unresolved as before.
Pinned by test_ambiguous_old_name_does_not_fall_through_to_the_relocation_alias.
self-added: yes

@bolichen97
bolichen97 force-pushed the skills/prepare-pr-slim branch from 8ba217e to 67e1a30 Compare October 4, 2026 11:17
@bolichen97

Copy link
Copy Markdown
Contributor Author

/ai-review override first-principles 67e1a30: agent: records maintainer @bolichen97's explicit decisions for #16731 - full loop arms auto-merge by default with a 24h watch, a writer's agent may override false-positive/over-engineering findings (never security-class) and must report it, red main gets its own T1 fix PR, the skill is renamed kirocrew-prepare-pr and scoped to this repo.

@github-actions github-actions Bot removed the readiness: checking Automated validation is still running label Oct 4, 2026
@bolichen97
bolichen97 force-pushed the skills/prepare-pr-slim branch from 92f9275 to b52e945 Compare October 4, 2026 13:04
@bolichen97

Copy link
Copy Markdown
Contributor Author

Rebased: main moved in docs/system-specs/modules/learn-cron-dashboard.md since this head went green (6878200 -> 5786c77); CI re-running.

@bolichen97

Copy link
Copy Markdown
Contributor Author

/ai-review override first-principles b52e945: agent: records maintainer @bolichen97's explicit decisions for #16731 - full loop arms auto-merge by default with a 24h watch, a writer's agent may override false-positive/over-engineering findings (never security-class) and must report it, red main gets its own T1 fix PR, the skill is renamed kirocrew-prepare-pr and scoped to this repo.

@github-actions github-actions Bot removed the readiness: passed Eligible automated validation passed for the current revision label Oct 4, 2026
@github-actions

github-actions Bot commented Oct 4, 2026

Copy link
Copy Markdown
Contributor

Human judgment recorded

@bolichen97 marked the first-principles AI finding as false positive, not applicable, or explicitly accepted for b52e9459ec92fce244328deff58d2ef90f2c44c8.

agent: records maintainer @bolichen97's explicit decisions for #16731 - full loop arms auto-merge by default with a 24h watch, a writer's agent may override false-positive/over-engineering findings (never security-class) and must report it, red main gets its own T1 fix PR, the skill is renamed kirocrew-prepare-pr and scoped to this repo.

This decision applies only to this commit. A new push requires a new judgment.

@github-actions github-actions Bot added the readiness: checking Automated validation is still running label Oct 4, 2026
@github-actions

github-actions Bot commented Oct 4, 2026

Copy link
Copy Markdown
Contributor

Human judgment recorded, but the following reviewer lane(s) could not be re-run automatically: First Principles Review. Re-run the lane's latest workflow run for b52e9459ec92fce244328deff58d2ef90f2c44c8 manually from the Actions tab.

The skill every Kiro Crew author's agent runs to ship a PR had drifted
from CI and sat at its 68000-byte budget. It now states the CI rules
for inside and outside authors (issue, triage, tiers, every blocking
lane, PR Hygiene's Pattern harvest and Reader lines, overrides, the
merge queue) and adds three loop rules: a red main gets its own T1
issue and fix PR while the original waits; the full loop arms
auto-merge and watches up to 24 hours, rebasing conflicts; and a
writer's agent may override a false positive or over-engineering
finding and must report it, while security-class findings always go
to a person.

It is renamed to kirocrew-prepare-pr everywhere it is referenced, and
carries `repo_scope: src/kiro_crew`, so it loads in every Kiro Crew
session and never in another repository's; the CRUX `cr` triggers are
gone. _RELOCATED_SKILLS quarantines an installed kirocrew-dev/prepare-pr
copy on upgrade, and a saved `$prepare-pr` still resolves to it.

The text shrinks from 67911 to about 55000 bytes and the ceiling test
ratchets to 56000. The CI doc's deferral paragraph now names the
tier-based path.

Closes #16731
@bolichen97
bolichen97 force-pushed the skills/prepare-pr-slim branch from b52e945 to 608b4aa Compare October 4, 2026 13:40
@bolichen97

Copy link
Copy Markdown
Contributor Author
  • span=8d3c83ce04bd fixed: Relocation renames a linked provider's manifest

Fixed in 608b4aa. _ensure_builtin_skills now checks every directory on the way to the old SKILL.md with is_link_or_junction and skips the quarantine with a warning when one is a link, so it never renames a file through a link or junction.
New test test_linked_old_dir_is_never_quarantined_through_the_link fails on the old code and passes now.

@bolichen97

Copy link
Copy Markdown
Contributor Author

/ai-review override first-principles 608b4aa: agent: records maintainer @bolichen97's explicit decisions for #16731 - full loop arms auto-merge by default with a 24h watch, a writer's agent may override false-positive/over-engineering findings (never security-class) and must report it, red main gets its own T1 fix PR, the skill is renamed kirocrew-prepare-pr and scoped to this repo.

@github-actions

github-actions Bot commented Oct 4, 2026

Copy link
Copy Markdown
Contributor

Human judgment recorded

@bolichen97 marked the first-principles AI finding as false positive, not applicable, or explicitly accepted for 608b4aa8acba5006ca8d5ac18fd867ee95ddcbc7.

agent: records maintainer @bolichen97's explicit decisions for #16731 - full loop arms auto-merge by default with a 24h watch, a writer's agent may override false-positive/over-engineering findings (never security-class) and must report it, red main gets its own T1 fix PR, the skill is renamed kirocrew-prepare-pr and scoped to this repo.

This decision applies only to this commit. A new push requires a new judgment.

@github-actions

github-actions Bot commented Oct 4, 2026

Copy link
Copy Markdown
Contributor

Human judgment recorded, but the following reviewer lane(s) could not be re-run automatically: First Principles Review. Re-run the lane's latest workflow run for 608b4aa8acba5006ca8d5ac18fd867ee95ddcbc7 manually from the Actions tab.

@bolichen97

Copy link
Copy Markdown
Contributor Author
  • span=5ab6f7c8de0d needs-a-decision: agent override and default auto-merge

Both rules are the maintainer's explicit instruction in #16731 (@bolichen97): arm auto-merge by default and let the agent override only false-positive or over-engineering findings, reported afterwards. Security, data-loss, corruption, crash and removed-guard findings still go to a person (the GPT data-loss finding on this PR was fixed, not overridden).
The PR cannot merge without a human approval from a reviewer other than the author (reviewDecision=REVIEW_REQUIRED), which is the "maintainer approves them on this PR" path.

@bolichen97

Copy link
Copy Markdown
Contributor Author
  • span=398b97dc67b2 needs-a-decision: path-mapped operators lose the skill silently

This is declared as Breaking in the PR body: an agent spec that names the old skill://…/kirocrew-dev/prepare-pr/SKILL.md path must be re-mapped on the crew page. Name-based lookups keep working through the relocation alias ($prepare-pr resolves to the new skill).
Whether to add a spec-rewriting migration instead is left to the maintainer; it is new scope for this PR.

@bolichen97

Copy link
Copy Markdown
Contributor Author
  • span=00b819c6e8a4 rebutted: bare "fix ci" falls through to the full loop

The full loop, including pushes and auto-merge, is the maintainer's chosen default for a Kiro Crew PR (#16731). The skill only loads under repo_scope: src/kiro_crew, so these triggers never fire outside this repo, and a user "hold" or a narrower request ("just fix the test") still disables auto-merge.

@github-actions github-actions Bot added readiness: passed Eligible automated validation passed for the current revision and removed readiness: checking Automated validation is still running merge conflict Branch has merge conflicts with its base — author must resolve before merge labels Oct 4, 2026
@chenmingwei23
chenmingwei23 merged commit 2359397 into main Oct 4, 2026
103 of 104 checks passed
@chenmingwei23
chenmingwei23 deleted the skills/prepare-pr-slim branch October 4, 2026 15:27
@github-actions github-actions Bot removed the readiness: passed Eligible automated validation passed for the current revision label Oct 4, 2026
bolichen97 added a commit that referenced this pull request Oct 5, 2026
PR Readiness reads main's lane list, which waits on an Issue Gate check.
Issue Gate is a pull_request workflow, so GitHub reads it from the PR's
merge ref, and release/0.8.0 has no issue-gate.yml. Every PR into the
release branch therefore waits on a check that never starts.

Copy .github/workflows/issue-gate.yml from main unchanged. Its grammar
script is read from a checkout of the default branch at run time, so the
workflow file is the only piece the release branch needs.

Source: origin/main 2359397 (file last changed there in #16740;
introduced by #16241, #16634, #16697).

Fixes #16868
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

prepare-pr skill: slim it down and bring it up to date with current CI

2 participants