fix(ci): pin skills feed deploy to production branch - #606
Merged
Conversation
Pin wrangler pages deploy to --branch=dev so release-tag builds (detached HEAD) stop landing on a preview alias while feeds.netclaw.dev stays frozen on the last dev-branch deploy. Add a post-deploy check that polls the custom domain until the manifest updatedAt matches the freshly generated one, so future routing regressions fail the workflow loudly instead of silently shipping a stale manifest.
3 tasks
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Fixes #591. The system skill feed has been serving a stale 2026-03-20 manifest for two releases, causing 5 of 6 skills to 404 on every daemon sync.
Root cause.
publish_skills.ymlrunswrangler pages deploy feeds/ --project-name=netclaw-feedswithout pinning a branch. On release-tag builds the runner is in detached HEAD, so wrangler infers a branch name ofheadand ships the new manifest to thehead.netclaw-feeds.pages.devpreview alias instead of the production custom domain. Meanwhilefeeds.netclaw.devhas been frozen on the last successfuldev-branch dispatch, advertising R2 keys for skills that were renamed/removed and never existed under those paths.Evidence during investigation:
head.netclaw-feeds.pages.dev/skills/.system/manifest.json→updatedAt2026-04-03, correct 5 skillsfeeds.netclaw.dev/skills/.system/manifest.json→updatedAt2026-03-20, pre-rename namesnetclaw-memory/1.1.0/SKILL.md→ 200,netclaw-diagnostics/0.6.0/SKILL.md→ 404Changes
--branch=devonwrangler pages deployso tag-triggered runs still hit the production branch regardless of git checkout state.feeds.netclaw.dev/skills/.system/manifest.jsonand asserts theupdatedAtmatches the freshly generated manifest. Fails loudly with a clear routing hint if it does not. The existing R2 spot-check only validatedskills.netclaw.dev, which is why this regression sailed through unnoticed.Operator action after merge
Once this lands, manually dispatch the workflow from
devto republish and unstick the live feed:Daemons will self-heal on their next
SystemSkillSyncServicerebuild.Follow-up
#605 tracks surfacing sync failures louder on the daemon side (doctor check, startup banner, stats) so the next time the feed breaks it is not buried in
daemon.log.Test plan
gh workflow run publish_skills.yml --ref devcurl https://feeds.netclaw.dev/skills/.system/manifest.jsonshows a freshupdatedAtand only the current 5 skill namesskill-authoringURL from the new manifest returns 200.sync-state.jsonpicks up all 5 skills without 404s