ci-failure-batching · zaxbyhub, generated — Claude Code skill: what it does, how to install
The ci-failure-batching skill blocks the agent from reporting tests as passing until the whole suite has run with the command the repo actually uses.
MIT ★ 539 updated 2026-10-29
npx skills add ZaxbyHub/opencode-swarm --skill ci-failure-batching -a claude-codeWhat ci-failure-batching does
One rule, no exceptions: no green report until the agent has run the entire suite the way the repository runs it. A single test file doesn't qualify.
The skill is a SKILL.md file produced by opencode-swarm's skill generator. It carries a delegation line — SKILLS: file:.opencode/skills/generated/ci-failure-batching/SKILL.md — that goes into a subagent task so the rule travels with the work.
On the review side, local test count and coverage get matched against the CI baseline, which means a run that quietly skipped half the suite shows up as a mismatch. Full runs take longer than the one file you were editing. That cost is the point, because CI runs the rest regardless.
What it helps with
- run the whole suite before calling tests passed
- match the local test count to the CI baseline
- hand the rule to a subagent through the
SKILLS:line - stop a single-file run from counting as green
Example prompts for the agent
- Run the full suite with the project's canonical command, then tell me whether it is green.
- Compare how many tests you ran and the coverage with the CI baseline, show me the gap.
- Put
SKILLS: file:.opencode/skills/generated/ci-failure-batching/SKILL.mdinto the subagent task.
How to install ci-failure-batching
1Skills CLI (Vercel)
npx skills add ZaxbyHub/opencode-swarm --skill ci-failure-batching -a claude-code2Manually: put the files in place
~/.claude/skills/ci-failure-batching/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/ci-failure-batching/
1Skills CLI (Vercel)
npx skills add ZaxbyHub/opencode-swarm --skill ci-failure-batching -a codex2Manually: put the files in place
.agents/skills/ci-failure-batching/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add ZaxbyHub/opencode-swarm --skill ci-failure-batching -a cursor2Manually: put the files in place
.cursor/skills/ci-failure-batching/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install ZaxbyHub/opencode-swarm .opencode/skills/generated/ci-failure-batching2Manually: put the files in place
.github/skills/ci-failure-batching/SKILL.md1Skills CLI (Vercel)
npx skills add ZaxbyHub/opencode-swarm --skill ci-failure-batching -a gemini-cli2Platform CLI
gemini skills install https://github.com/ZaxbyHub/opencode-swarm --path .opencode/skills/generated/ci-failure-batchingManually: put the files in place
.devin/skills/ci-failure-batching/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add ZaxbyHub/opencode-swarm --skill ci-failure-batching -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about ci-failure-batching
How do I install ci-failure-batching in Claude Code?
Run npx skills add ZaxbyHub/opencode-swarm --skill ci-failure-batching -a claude-code. The skill file lands at ~/.claude/skills/ci-failure-batching/SKILL.md.
What does ci-failure-batching do?
It stops the agent from declaring tests passed after a partial run: the full suite has to go through the project's canonical command first.
What does the reviewer check look at?
It puts your local test count and coverage next to the CI baseline. A gap means part of the suite never ran.
Source text
SKILL.md · MIT
--- name: ci-failure-batching description: ci-failure-batching metadata: generated_from_knowledge: '["b96cf48c-9bfa-4ab2-9ef1-0128f48bbfd7"]' source_knowledge_ids: '["b96cf48c-9bfa-4ab2-9ef1-0128f48bbfd7"]' generated_at: "2026-07-14T05:31:51.670Z" confidence: "0.73" status: active version: "2" skill_origin: generated --- <!-- generated by opencode-swarm skill-generator. Do not edit by hand; edits will be preserved on regeneration only with controlled update mode. --> # ci-failure-batching ## Trigger - (no explicit trigger metadata; cluster derived from category/tags) ## Required Procedure - run the project's full test suite using the canonical command before declaring tests pass ## Forbidden Shortcuts - (none recorded) ## Delegation Template When delegating a task affected by this skill, include: ``` SKILLS: file:.opencode/skills/generated/ci-failure-batching/SKILL.md ``` ## Reviewer Checks - compare local test count/coverage against CI baseline ## Source Knowledge IDs - b96cf48c-9bfa-4ab2-9ef1-0128f48bbfd7 — Always run the project's canonical test command (e.g. bun test) rather than a subset; partial runs may mask failures that CI catches.
© ZaxbyHub · MIT · ZaxbyHub/opencode-swarm
Similar skills
All skills for Testing and QA →ci-failure-batching by other authors comes first.