fix(T12094): grant actions:write so the CI dispatch can actually fire - #1185
Merged
Conversation
The dispatch step shipped in #1182 was correct but could not run. Measured on the v2026.8.5 release: could not create workflow dispatch event: HTTP 403: Resource not accessible by integration Creating a `workflow_dispatch` event requires `actions: write`, and the workflow declared only `contents`, `pull-requests` and `id-token`. So the bump-PR still opened with no checks — the exact failure the step exists to prevent. The step's graceful degradation earned its keep: because it warns rather than fails, the release completed and named the cause precisely instead of dying at the last step with a bare non-zero exit. Verified: both files now parse with `actions: write`, and template parity passes in `--strict`. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
This is the last thing blocking an unattended release, and it is the same root
cause one layer down.
v2026.8.5's bump-PR was BLOCKED by the Worktree NAPI Prebuild Gate:
Found prebuild run #31456705101 (status: completed)
✗ worktree-napi prebuild completed with conclusion: action_required
`action_required` does not mean the build failed. It means GitHub queued the run
pending human approval — which is the state of EVERY run triggered by a
GITHUB_TOKEN-created `pull_request`, i.e. every release bump-PR. The gate treated
any non-`success` conclusion as failure, so it reported a build failure for a
build that had never been permitted to start.
Two changes, both addressing the cause rather than the symptom:
1. The gate now recognises `action_required`, APPROVES the run, and waits for the
real result. `actions: write` is granted on that job so it can. If the token
still cannot approve, it prints the exact `gh api -X POST …/approve` command
instead of a misleading failure.
2. `release-prepare` now dispatches `worktree-napi-prebuild.yml` for the release
branch alongside `ci.yml`. An explicit dispatch is not subject to the approval
gate, so the prebuild produces a real conclusion in the first place and the
approval path becomes a fallback rather than the norm.
Verified: `ci.yml` parses, the gate job carries `actions: write`, template parity
passes `--strict`, and gate 15 confirms every `cleo` verb in the workflows exists.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The dispatch step from #1182 was correct but couldn't run. Measured on the v2026.8.5 release:
Creating a
workflow_dispatchevent needsactions: write; the workflow declared onlycontents,pull-requests,id-token. So the bump-PR still opened with zero checks.Worth noting the step's graceful degradation paid off: because it warns instead of failing, the release completed and told us exactly why, rather than dying at the last step with a bare non-zero exit.
Template parity passes in
--strict.🤖 Generated with Claude Code