lidge-jun/codexclaw / plugins/codexclaw/skills/dev
cxc-dev
VerifiedUse for coding, PR delivery, scaffolding and QA. Triggers: develop, fix, refactor, test, review, docs, browse, stacked PR, 개발, 수정, 검토, 스택 PR.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 12.0 KB |
| agents/openai.yaml | config | 143 B |
| references/async-questions.md | markdown | 6.2 KB |
| references/browser-routing.md | markdown | 3.4 KB |
| references/code-mode-examples.md | markdown | 4.3 KB |
| references/development-practice.md | markdown | 6.3 KB |
| references/hosted-ci-evidence.md | markdown | 3.2 KB |
| references/logging.md | markdown | 2.8 KB |
| references/methodology-overlays.md | markdown | 2.8 KB |
| references/native-execution.md | markdown | 6.3 KB |
| references/peer-collaboration.md | markdown | 10.6 KB |
| references/product/crud-product-development.md | markdown | 2.7 KB |
| references/reader-documents.md | markdown | 6.7 KB |
| references/skill-catalog.md | markdown | 14.1 KB |
| references/skill-ownership.md | markdown | 5.0 KB |
| references/stacked-prs.md | markdown | 16.0 KB |
| references/static-analysis-gate.md | markdown | 2.6 KB |
| references/static-analysis.md | markdown | 3.3 KB |
references/hosted-ci-evidence.md
View on GitHub# Hosted CI evidence
### Hosted CI evidence (DEV-CI-EVIDENCE-01, DEFAULT)
Zero failing checks is not the same as passing tests. Hosted CI has several ways to
report nothing while looking like an answer, and each one reads as green:
1. **Required jobs never started.** A rollup filtered for failures returns an empty list
when only lightweight checks ran. Fork-approval gating is one possible cause, not an
automatic diagnosis — investigate rather than conclude.
2. **An aggregate fails with no failing test.** Executed legs succeed, a dependency is
cancelled or skipped, and the aggregate rejects the run. Cancellation is missing
completion evidence, not proof of a defect.
3. **A partial rerun regenerates nothing.** `gh run rerun --failed` requests
`rerun-failed-jobs`; its zero exit means the request was accepted, not that the
missing evidence now exists.
4. **Exit zero while pending or cancelled.** `gh run view --exit-status` keys on failure
states, so a cancelled or still-pending conclusion exits 0. It is not a success
predicate.
5. **Evidence from the wrong event.** Selecting the newest run for a SHA can return a
manual dispatch. A matching SHA does not establish equivalent coverage.
So before claiming hosted CI passed: confirm the expected jobs actually ran, and identify
the PR head, the SHA the checks really ran against, the workflow event, the run id and
the attempt. Distinguish approval-blocked, absent, skipped, cancelled, pending and failed
— they need different responses. Inspect an aggregate's dependencies rather than its
headline. Approving a workflow or triggering a rerun is a separate authorization from
diagnosing one, and a full rerun is warranted only when inspection shows a partial rerun
cannot regenerate what is missing.
```sh
gh pr view <PR> -R <REPO> --json headRefOid,statusCheckRollup
gh api --paginate 'repos/<REPO>/commits/<HEAD_SHA>/check-runs?per_page=100'
gh run list -R <REPO> --commit <HEAD_SHA> \
--json databaseId,event,headSha,status,conclusion,workflowName
gh run view <RUN> -R <REPO> --json event,headSha,attempt,status,conclusion,jobs
```
Compare the returned jobs against what the workflow should have produced, and correlate
merge-ref checks where they apply. An empty check-run list is not proof of why the checks
are absent.
**Reading a completed job's log while its workflow is still running.** `gh run view
--job <JOB> --log` checks that the parent RUN finished, not just the job, so it refuses a
job that is already complete inside a run that is not. Fetch the job log endpoint
directly instead, and read the command's status, stderr and returned content before
concluding the logs are unavailable:
```sh
gh api 'repos/<REPO>/actions/jobs/<JOB>' --jq '{id,run_id,head_sha,status,conclusion}'
gh api 'repos/<REPO>/actions/jobs/<JOB>/logs'
```
If that raw response is rejected for escape sequences, newer CLI versions expose
`--allow-escape-sequences`. Treat it as conditional, not standard: it does not exist in
every installed version, and it relaxes a deliberate protection — read the result through
a non-executing text reader rather than printing it to a terminal. Keep the run id,
attempt and job id when comparing reruns, and remember old logs are subject to retention.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| finishing-a-development-branchobra/superpowersUse when implementation is complete, all tests pass, and you need to decide how to integrate the work | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| systematic-debuggingobra/superpowersUse when encountering any bug, test failure, or unexpected behavior, before proposing fixes | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| test-driven-developmentobra/superpowersUse when implementing any feature or bugfix, before writing implementation code | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| skill-creatoranthropics/skillsCreate new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |