Repository navigation
Closed
Labels
agent-readyFully specified and briefed; eligible for autonomous pickup from the frontier.Fully specified and briefed; eligible for autonomous pickup from the frontier.
Description
Parent
Refs #15703
Source: PLAN Phase 8.5, topic ecosystem-specifics
Depends on
Outcome
The repository named in this item's title resolves its build, test and lint commands from its own ecosystem records, so its runs do not report INCOMPLETE once unattended runs stop using the marketplace's bundled defaults.
Demo: on that repository's default branch, .claude/ecosystems/ holds its records, and a /toolchain:check all run there shows no skip (unresolved: ...) cell.
The target repository belongs to the owner (melodic-software). It is one of 15 that have a Claude lane workflow and no ecosystem records, counted on 2026-11-12. The pull request is opened in that repository; this item tracks it.
Steps, the same as for this repository's own records (#15704):
- Pick one merged commit range in the target repository that touches at least one file of each ecosystem it has; record both SHAs in the pull request body.
- Run
/toolchain:check alland/toolchain:lintthere with the bundled defaults in place. Take the baseline command list from the Bash commands the run executed and put it in the pull request body. - Write one ecosystem record per covered ecosystem, following that repository's own rules where a bundled default differs, each difference listed as intended. Validate each record against the ecosystem-commands schema.
- Re-run on the same diff; the command list matches the baseline apart from the listed intended differences.
- Shell, markdown, YAML and PowerShell have no build file to infer a command from, so each of those the repository uses needs a record.
- Run budget: 6 runs; launch nothing after a usage-limit error.
- Check the diff for secrets and machine-specific data before the first push.
Decisions
- Owner, 2026-11-12: "I don't want anything to be 'we cannot accept incomplete.' We'll need to run that setup in each repository. We'll need to make sure that everything is set up and in place."
- D7: records land in the owner's other lane repositories before unattended runs change.
- D2: records land by pull request; nobody types
/toolchain:setup apply. If the target repository turns out to need the user-typed setup, stop and give the owner the one command. - Owner, 2026-11-12 (scope): every owner repository with a Claude lane workflow and no records is covered, 15 in all.
Key interfaces
- The record's format is the existing ecosystem-commands schema.
Done when
- AC5 (precondition): the target repository has a merged pull request adding its records, and
.claude/ecosystems/on its default branch is non-empty. - Each record validates against the ecosystem-commands schema.
- A recorded
/toolchain:check alloutput for that repository contains noskip (unresolved. - The merged pull request's URL is posted on this item.
Out of scope
- Any change to the target repository other than its ecosystem records.
- Any repository the owner does not own.
Metadata
Metadata
Assignees
Labels
agent-readyFully specified and briefed; eligible for autonomous pickup from the frontier.Fully specified and briefed; eligible for autonomous pickup from the frontier.
Type
Fields
Work status
confirmed
Priority
medium
Activity
melodic-automation-lanes commentedon Nov 13, 2026
No work-resolve-issue run started in the 1 hour since
agent-readywas last added. work-unblock removed and addedagent-readyback to start a fresh run on main's workflow files. If this run ends the same way, it re-fires the issue again, waiting longer each time.github-actions commentedon Nov 13, 2026
The triage sweep dispatched intake triage for this issue (dispatch 1; 2 runs that reach intake, or 6 dispatches, hand it to a maintainer).
github-actions commentedon Nov 13, 2026
Size: Small
Complexity: Medium
Uncertainty: Knowable
Flags: none
Blocked by: none
Understood: add ecosystem records under
.claude/ecosystems/in the owner'sstandardsrepository by pull request, so/toolchain:check allthere shows noskip (unresolved: ...)cell. Baseline and re-run evidence go in that pull request's body.The outcome, steps, decisions and done-when are exact, so I'm asking nothing further. It is one of 15 sibling items for the other owner repositories, each covering a different repository, so I'm not treating it as a duplicate. The dependency #15704 is not in the open issue list. No merged pull request already fixes this.
Automated triage. A maintainer reviews it.
melodic-automation-lanes commentedon Nov 13, 2026
Not admitted to the issue lane: its body or a criterion names .claude/ecosystems/, which the issue lane cannot push (patch-protected-path); a builder session takes it; it carries host-run. Admission takes the issue once each of those is met, when intake triages it again (
.github/actions/resolve-config/admission-rules.json).chore(ecosystems): add this repository's ecosystem records (#1002)
fix: fail the multi-command check-cmd records on any failed check (#1007
fix(ecosystems): follow CI's current actionlint fork and gitleaks sca…
kyle-sexton commentedon Nov 16, 2026
Delivered; closing.
gitleaks dir . --config .gitleaks.toml --no-banner --redactat 8.30.1, and the composite pins for typos, editorconfig-checker and zizmor./toolchain:check alloutput: in docs(disk-hygiene): split Windows audit/execution lanes, scope the guard wording, fix stale no-userConfig claim #1031's Verification section. It holds noskip (unresolvedcell; all seven selected ecosystems resolved from the repository's records. Its overall line reads FAIL on two tool-version differences of the machine that ran it (golangci-lint 2.15.0 against the 2.12.2 pin, and a ruff other than CI's 0.15.14); the records keep CI's pins.