agentleFS
HeyEddi-comHeyEddi-com/heyeddi-skills/skills/heyeddi-ci-runners/SKILL.md

heyeddi-ci-runners

Skill0 starsChanged 3 months agoMITView on GitHub

Configure HeyEddi Spot runners — prefer GHA overflow/always with the HEYEDDI_RUNS_ON hook; author sealed eddi-ci.yaml pipeline jobs only for no-GHA or niche checks. Use when wiring runners, pipeline YAML, or Spot CI.

The file

We split this skill into the rules it gives an agent

9 rules. About 0 are common, 6 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈6

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  • could be automatic: a setting could enforce it
  1. Very specific:

    Golden rule: Never duplicate the same test/build in GHA and pipeline:.

    line 20
  2. Distinctive:

    Ensure every Linux job uses:

    line 27
  3. Very specific:

    Keep eddi-ci.yaml Reviewer-only with pipeline: {} (use @heyeddi-ci-config).

    line 37
  4. Very specific:

    Prefer @heyeddi-ci-config for the create/update decision gate.

    line 48
  5. Distinctive:

    Before any summary that implies execution, run:

    line 55
  6. Distinctive:

    Never secrets in YAML

    line 81
  7. Distinctive:

    Never invent test commands

    line 82
  8. Distinctive:

    Never merge without authorize merge

    line 84
  9. Distinctive:

    Never claim correlated AI+runner timeline (Coming soon)

    line 85

≈ “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

HeyEddi CI Runners

Spot is shipped. Help the user wire the right runner path and never invent execution claims.

Two job kinds (do not collapse)

KindWhereCheck nameWhen to use
GHA on Spot.github/workflows/*.yml + workspace gha_spot_modeSame workflow CheckRepos that already have Actions (default)
Sealed pipeline:eddi-ci.yaml pipeline: jobsHeyEddi Runner: {job_id}No (or minimal) GHA, or niche sealed checks

Golden rule: Never duplicate the same test/build in GHA and pipeline:. Prefer GHA + Spot routing; keep pipeline: {}.

Product docs: https://ci.heyeddi.com/docs · hub mirrors: see reference/runners.md.

Prefer GHA overflow / always

  1. Confirm .github/workflows/ exists (or user wants Actions-based CI).
  2. Ensure every Linux job uses:
runs-on: ${{ vars.HEYEDDI_RUNS_ON || 'ubuntu-latest' }}
  1. Tell the user to set workspace Runners mode on ci.heyeddi.com:
    • overflow — GitHub minutes first; Spot JIT when exhausted
    • always — Linux jobs on Spot from the first run
    • off — hosted only
  2. Keep eddi-ci.yaml Reviewer-only with pipeline: {} (use @heyeddi-ci-config).
  3. Limits: Linux only; private repos; prepaid/entitled workspace; platform kill switch is ops-owned. Overflow reliability can still gap — do not promise “always works.”

Sealed pipeline: (advanced / no-GHA)

  1. load_policy_contract — living knobs only.
  2. inspect_repo — evidenced commands + path prefixes; never invent pytest / npm test.
  3. Explicit user opt-in (billable runner time).
  4. Author argv-safe run (no shell operators, sudo, curl, wget), catalog image, filter.paths when possible.
  5. Jobs run on Spot when the workspace is entitled and Spot env is live; they post HeyEddi Runner: {job_id} Checks.

Prefer @heyeddi-ci-config for the create/update decision gate.

Evidence rule (mandatory)

  • Jobs can run under entitlement + Spot substrate.
  • A job did run only when you have Check / Spot evidence (HeyEddi Runner: …, workflow Check conclusion, Ops/logs).
  • YAML lint ≠ execution. Unknown → say so and use @heyeddi-ci-fails.
  • Before any summary that implies execution, run:
assert_runners_claims --check --agent-text "<your summary>"

Pipeline

load_policy_contract
inspect_repo
→ choose GHA-first vs sealed pipeline (never both for the same CI)
→ wire HEYEDDI_RUNS_ON hook and/or author pipeline: jobs
assert_runners_claims --check --agent-text "<your summary>"

Tools

ToolPurpose
load_policy_contractLiving knobs / guide
inspect_repoLanguages + evidenced commands
assert_runners_claimsBlock unsubstantiated execution claims

Safety

  • Never secrets in YAML
  • Never invent test commands
  • Never auto_merge
  • Never merge without authorize merge
  • Never claim correlated AI+runner timeline (Coming soon)

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.