agentleFS
ldaumeldaume/agentic-engineering-harness/skills/infrastructure/gitea-actions/SKILL.md

gitea-actions

Skill3 starsChanged 3 months agoMITView on GitHub

Designs, debugs, secures, and reviews Gitea Actions workflows, act_runner setups, runner labels, tokens, secrets, variables, cache, artifacts, and GitHub Actions compatibility. Use when working with .gitea/workflows, Gitea CI, act_runner, GITEA_TOKEN permissions, self-hosted runners, workflow migration, sparse-checkout CI helpers, docker/build-push-action on Gitea, step summaries, or Gitea-specific CI failures.

The file

Show its rules and text

We split this skill into the rules it gives an agent

22 rules. About 0 are common, 15 are distinctive, and 7 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈15

    Distinctive · estimate

    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
  1. Distinctive:

    Use this for Gitea-hosted CI/CD, especially when GitHub Actions examples need to be adapted to Gitea's runner, token, and network model.

    line 8
  2. Very specific:

    Read AGENTS.md, existing .gitea/workflows/*.yml, runner docs, deployment docs, and repo variables/secrets conventions.

    line 15
  3. Distinctive:

    Verify which GitHub Actions syntax is actually supported by the installed Gitea version before copying examples.

    line 19
  4. Distinctive:

    Keep secrets, registry credentials, and deploy tokens scoped to the smallest workflow that needs them.

    line 21
  5. Distinctive:

    If current behavior depends on self-hosted network topology, document the assumption in the workflow or agent docs.

    line 23
  6. Very specific:

    Put workflows under .gitea/workflows/.

    line 28
  7. Very specific:

    Prefer explicit on: triggers and path filters for expensive jobs.

    line 29
  8. Distinctive:

    Keep preflight, verify, build, image, and deploy jobs separate enough that failures point to a real layer.

    line 32
  9. Distinctive:

    Use concurrency/cancel-in-progress when repeated pushes to the same ref should not burn runner capacity.

    line 34
  10. Distinctive:

    Keep cache keys tied to lockfiles and relevant runtime versions.

    line 36
  11. Very specific:

    After any automated workflow mutation (fan-out, codegen, bulk step inject), parse every touched YAML file before push.

    line 38
  12. Very specific:

    If a job uses sparse-checkout, include every path that job invokes.

    line 41
  13. Very specific:

    If you keep docker/build-push-action, set job env DOCKER_BUILD_RECORD_UPLOAD=false and DOCKER_BUILD_SUMMARY=false so the Complete job does …

    line 47
  14. Distinctive:

    Avoid privileged runners by default.

    line 57
  15. Distinctive:

    Treat runner filesystem state as disposable unless explicitly managed.

    line 58
  16. Distinctive:

    Document any required host mounts or Docker-in-Docker assumptions.

    line 59
  17. Very specific:

    Use GITEA_TOKEN only for the permissions the workflow needs.

    line 63
  18. Distinctive:

    Prefer repo or org variables for non-secret configuration.

    line 64
  19. Distinctive:

    Never echo tokens, Docker auth JSON, private keys, or provider credentials.

    line 65
  20. Distinctive:

    Check unsupported syntax first: permissions, reusable workflows, services, cache/action versions, OIDC, environments, and marketplace …

    line 74
  21. Distinctive:

    Test on a small branch before wiring deploy jobs.

    line 78
  22. Distinctive:

    Keep the workflow readable.

    line 79

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Gitea Actions

Use this for Gitea-hosted CI/CD, especially when GitHub Actions examples need to be adapted to Gitea's runner, token, and network model.

Project Fit Check

Before changing workflows:

  1. Read AGENTS.md, existing .gitea/workflows/*.yml, runner docs, deployment docs, and repo variables/secrets conventions.
  2. Detect the Gitea version, act_runner version, runner labels, container mode, and whether jobs run on Docker, host shell, or custom images.
  3. Verify which GitHub Actions syntax is actually supported by the installed Gitea version before copying examples.
  4. Keep secrets, registry credentials, and deploy tokens scoped to the smallest workflow that needs them.
  5. If current behavior depends on self-hosted network topology, document the assumption in the workflow or agent docs.

Workflow Rules

  • Put workflows under .gitea/workflows/.
  • Prefer explicit on: triggers and path filters for expensive jobs.
  • Pin third-party actions when possible; avoid floating major tags for sensitive deploy paths.
  • Keep preflight, verify, build, image, and deploy jobs separate enough that failures point to a real layer.
  • Use concurrency/cancel-in-progress when repeated pushes to the same ref should not burn runner capacity.
  • Keep cache keys tied to lockfiles and relevant runtime versions.
  • Upload artifacts only when they help debugging or deployment.
  • After any automated workflow mutation (fan-out, codegen, bulk step inject), parse every touched YAML file before push. Insert steps under an existing steps: list with the same indent as siblings.
  • If a job uses sparse-checkout, include every path that job invokes. Prefer a helper directory pattern (for example scripts/ci/) over one-file lists.
  • Soft-skip writing a job summary when GITHUB_STEP_SUMMARY is unset (local runs). Do not soft-fail summary or other always() helper steps in CI to hide a missing sparse path; fix the checkout list instead.
  • On self-hosted Gitea, prefer raw docker / docker buildx for image builds. If you keep docker/build-push-action, set job env DOCKER_BUILD_RECORD_UPLOAD=false and DOCKER_BUILD_SUMMARY=false so the Complete job does not upload .dockerbuild records via the GitHub Artifact API.

Runner Rules

  • Match runs-on labels to registered runner labels exactly.
  • Know whether job containers can reach the Gitea host, package registry, Docker socket, private networks, and deployment targets.
  • Avoid privileged runners by default.
  • Treat runner filesystem state as disposable unless explicitly managed.
  • Document any required host mounts or Docker-in-Docker assumptions.

Token And Secret Rules

  • Use GITEA_TOKEN only for the permissions the workflow needs.
  • Prefer repo or org variables for non-secret configuration.
  • Never echo tokens, Docker auth JSON, private keys, or provider credentials.
  • For private package/container registries, use one clear auth path and test it in the job image that actually pulls or pushes.
  • Rotate or invalidate deploy tokens when workflow scope changes.

Migration From GitHub Actions

When porting:

  1. Check unsupported syntax first: permissions, reusable workflows, services, cache/action versions, OIDC, environments, and marketplace actions.
  2. Replace GitHub-specific APIs, URLs, and token assumptions.
  3. Validate action sources and whether the runner can fetch them.
  4. Test on a small branch before wiring deploy jobs.
  5. Keep the workflow readable; do not recreate GitHub-specific complexity if Gitea offers a simpler local path.

Verification

  • yamllint or parser check when available
  • dry-run or test branch workflow for new CI paths
  • one failing-path check for secrets/registry/network assumptions when feasible
  • git diff review for accidental secret insertion

Red Flags

  • copied GitHub workflow with no Gitea compatibility review
  • broad token permissions for read-only jobs
  • deploy job triggered on every branch
  • runner label that exists only on one machine but is undocumented
  • cache key that ignores lockfiles
  • private registry auth configured in one container but used in another
  • docker/build-push-action on self-hosted Gitea without DOCKER_BUILD_RECORD_UPLOAD=false (Complete-job CreateArtifact timeouts after a successful registry push)
  • sparse-checkout file lists that omit scripts or helpers the job still runs (exit 127 after an otherwise successful deploy or verify)
  • workflow YAML left unparsed after automated step injection (de-indented steps at column 0 that break Actions parse only after merge)
  • registry retention that deletes the sha256: child manifests behind a multi-arch (OCI index) job-image tag, or trims tags by count or age without checking every consumer repository's pins (the tag still lists in the package UI, but pulls fail with manifest unknown; runners with a cached image keep working until their cache is pruned)

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.