ci-cd-pipeline
Skill1 starsChanged 9 months agoMITView on GitHub
Analyze a repo and maintain its GitHub Actions CI/CD pipeline. Detects language, tooling, test frameworks, Docker, and deploy targets, then adds or removes workflow actions to match the project's current state. Use when the user asks to "add CI/CD", "update CI", "review pipeline", "set up GitHub Actions", "audit CI", "improve CI/CD", or any request about CI/CD pipelines, GitHub Actions workflows, or continuous integration.
- Reads credentials
9 other projects use a version of this.
Adopted by
Most recently taken up by
- Cure-Consulting-Group/ProductEngineeringSkillstheir own version, first seen Nov 12, 2026
- nisarth/claude-skills-storetheir own version, first seen Nov 8, 2026
- Wade-DevCode/awesome-coding-skills-cntheir own version, first seen Nov 8, 2026
- fideguch/my_pm_toolstheir own version, first seen Nov 1, 2026
- Sagargupta16/agent-recipestheir own version, first seen Nov 1, 2026
- nanvix/nanvixtheir own version, first seen Oct 8, 2026
Used with
Projects with it also use performance-review and api-design.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
AnthropicSkillperformance-review
Structure a performance review with self-assessment, manager template, and calibration prep. Use when review season kicks off and you need a self-assessment template, writing a manager review for a direct report, prepping rating distributions and promotion cases for calibration, or turning vague feedback into specific behavioral examples.
Also in 3 projects that use this
MicrosoftSkillapi-design
Guide for designing and documenting RESTful APIs. Use when asked to design an API, create endpoints, or document an API.
Also in 3 projects that use this
Other versions9 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 9
- nanvix/nanvix★ 319same name
- rloisell/rl-agents-n-skills★ 8same name
- Wade-DevCode/awesome-coding-skills-cn★ 7same name
- Sagargupta16/agent-recipes★ 4same name
- Cure-Consulting-Group/ProductEngineeringSkills★ 3same name
- fideguch/my_pm_tools★ 1same name
- impeto-ai/claude-skills-impeto★ 1same name
- nisarth/claude-skills-store★ 1same name
- Ygohel18/skills★ 1same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository9 other files it gives its agents.
- bug-bash-updateSkill
- claudemd-auditorSkill
- docs-consolidatorSkill
- ralphSkill
- refactor-auditSkill
- reflectSkill
- setup-workflowSkill
- skill-reviewerSkill
- smoke-testSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
14 rules. 0 are common, 7 are distinctive, and 7 are very specific to this project.
0
Common
Most of the 7 edited versions we read keep these.
7
Distinctive
This author's own choices. Worth a look.
7
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 87
When adding a deploy action, read ~/.claude/skills/ci-cd-pipeline/references/deploy-prerequisites.md for the detected platform and include …
- Distinctive: line 89
If running interactively, wait for user approval before making changes.
- Very specific: line 95
Create or edit .github/workflows/*.yml files
- Distinctive: line 96
If a new tool is needed (e.g., adding mypy job but mypy isn't in deps), add it to dev dependencies
- Very specific: line 97
If a tool config is missing (e.g., [tool.ruff] section), add it to the project config file
- Distinctive: line 98
Delete workflow files or jobs that are no longer needed
- Distinctive: line 99
Run the tools locally to verify the pipeline starts green (lint, typecheck, test)
- Distinctive: line 104
Prefer fewer workflow files with multiple jobs over many single-job files.
- Very specific: line 106
All jobs in ci.yml should run in parallel unless they have dependencies.
- Very specific: line 107
Use actions/checkout@v4 and actions/setup-python@v5 / actions/setup-node@v4.
- Very specific: line 110
When adding tooling config, use the project's config file (e.g., pyproject.toml for Python, package.json for JS).
- Distinctive: line 111
When removing an action, also clean up any orphaned tool configs that were only used by that action.
- Very specific: line 112
When proposing a CI job, derive the exact commands from the project's own config (package.json scripts, Makefile targets, existing dev …
- Distinctive: line 114
Don't propose actions for tools the project doesn't use.
“Common” is measured: kept by at least 6 in 10 of the 7 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
CI/CD Pipeline
Analyze a repository and maintain its GitHub Actions CI/CD pipeline. Designed to be called repeatedly — each invocation audits the current state and proposes additions and removals.
Workflow
Phase 1: Discover project state
Build a comprehensive project profile by scanning the entire repository:
- Language & runtime — Check file extensions, config files (
pyproject.toml,package.json,go.mod,Cargo.toml,Gemfile,foundry.toml,hardhat.config.*,truffle-config.js,Move.toml, etc.) - Package manager — pip/uv/poetry, npm/yarn/pnpm, cargo, go modules, etc.
- Tooling configs — Lint, format, typecheck tools. Don't assume defaults — read actual config files (
.eslintrc*,prettierin deps,ruffin deps,tsconfig.json,foundry.toml,biome.json, etc.) - Test framework — Detect from config files, dependency lists, and test file patterns (
tests/,test/,*_test.go,*.test.ts,*.t.sol, etc.) - Build tools — Read
package.jsonscripts,Makefiletargets,scripts/directory,foundry.toml,hardhat.config.*,Cargo.toml,pyproject.tomlbuild sections - Docker —
Dockerfile,docker-compose.yml,.dockerignore - Deploy targets — Railway, Fly.io, Vercel, AWS, Kubernetes manifests, Terraform, etc.
- Existing workflows — Read all
.github/workflows/*.ymlfiles. For each, note the jobs, triggers, and tools used. - Version requirements — From
requires-python,engines,.python-version,.nvmrc,.node-version - Dev dependencies — Check what's available in dev/test dependency groups
- Monorepo structure — Detect package boundaries (
packages/,apps/, workspace configs inpackage.json,pnpm-workspace.yaml, Cargo workspaces, etc.) - Custom scripts — Read
scripts/directory andpackage.jsonscripts to understand project-specific build/test/deploy commands
Phase 2: Reason about pipeline coverage
Based on the discovered project profile, determine what CI jobs are needed. Do NOT reference a static catalog — reason from what the project actually uses.
Evaluate coverage across these categories:
- Code quality — For each language/package detected, check if lint, format, and typecheck tools are configured. Identify the specific tool and command from the project's own config (e.g.,
pnpm lintfrompackage.jsonscripts,forge fmt --checkfromfoundry.toml,ruff check .frompyproject.toml). - Testing — For each test framework detected, identify the correct test command. Check for unit tests, integration tests, E2E tests, and fuzz tests. Look at
package.jsonscripts,Makefiletargets, and test config files to find the exact commands. - Security — Dependency audits based on detected package managers (
npm audit,pip-audit,cargo audit, etc.). Secret scanning if the project handles credentials or has.envfiles. - Build — Compilation steps based on detected build tools (
forge build,pnpm build,cargo build,go build,docker build, etc.). Only include if the project has build artifacts. - Deploy — Platform-specific deploy jobs based on detected deploy targets (Railway, Fly.io, Vercel, etc.).
For each proposed job:
- Derive the exact commands from the project's own config files — don't assume default commands.
- For monorepo/multi-package projects, determine if per-package jobs or matrix jobs are appropriate.
Compare against existing workflows to classify each item:
| Needed? | Exists in workflows? | Decision |
| Yes | No | Add |
| Yes | Yes, but stale/misconfigured | Update (explain what changed) |
| No | Yes | Remove (tooling no longer present) |
| Yes | Yes, correctly configured | Keep |
Phase 3: Present the plan
Present findings to the user:
## CI/CD Audit Report
### Actions to Add
- [ ] [action]: [rationale — what config/files were detected that justify this]
### Manual Setup Required
> Only include this section when adding a deploy action.
[Platform name]:
1. [step from deploy-prerequisites.md]
2. [step from deploy-prerequisites.md]
3. Add `SECRET_NAME` to GitHub repo secrets (Settings → Secrets and variables → Actions)
4. [verification step]
### Actions to Remove
- [ ] [action]: [rationale — signal no longer present]
### Actions to Update
- [ ] [action]: [what changed and why]
### No Changes Needed
- [action]: correctly configuredWhen adding a deploy action, read ~/.claude/skills/ci-cd-pipeline/references/deploy-prerequisites.md for the detected platform and include its setup steps in the Manual Setup Required section. This ensures the user knows what manual steps are needed before the workflow will function.
If running interactively, wait for user approval before making changes. If running autonomously (e.g., as a post-task audit subagent), proceed directly to Phase 4 — apply all additions and updates from the audit.
Phase 4: Execute changes
After approval:
- Create or edit
.github/workflows/*.ymlfiles - If a new tool is needed (e.g., adding mypy job but mypy isn't in deps), add it to dev dependencies
- If a tool config is missing (e.g.,
[tool.ruff]section), add it to the project config file - Delete workflow files or jobs that are no longer needed
- Run the tools locally to verify the pipeline starts green (lint, typecheck, test)
- Present a summary of all changes made
Guidelines
- Prefer fewer workflow files with multiple jobs over many single-job files.
- Standard layout:
ci.ymlfor lint/typecheck/test,security.ymlfor audits/scanning,deploy.ymlfor deployment. - All jobs in
ci.ymlshould run in parallel unless they have dependencies. - Use
actions/checkout@v4andactions/setup-python@v5/actions/setup-node@v4. - Pin action versions to major tags (e.g.,
@v4), not SHAs. - CI triggers:
pushto main/master +pull_request. Security:pushto main + weeklyschedule. - When adding tooling config, use the project's config file (e.g.,
pyproject.tomlfor Python,package.jsonfor JS). - When removing an action, also clean up any orphaned tool configs that were only used by that action.
- When proposing a CI job, derive the exact commands from the project's own config (package.json scripts, Makefile targets, existing dev scripts) rather than assuming default commands.
- For monorepo/multi-package projects, detect package boundaries and create per-package or matrix jobs as appropriate.
- Don't propose actions for tools the project doesn't use. Only propose what the project profile supports.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

