CI Validation
Testing & Qualityci-validationCI Validation is an Agent Skill for testing & quality published in the camunda/camunda repository on GitHub: a SKILL.md instruction file an AI agent can load to do this task.
Validates GitHub Actions workflow changes in this monorepo using actionlint, conftest policy checks, spotless formatting, and act testability assessment. Use when creating, editing, or reviewing workflow files.
How to use this skill
Bring this guide into your coding agent with a prompt tailored to the tool you use.
- Open your project in Codex.
- Copy and paste the prompt below.
- Review files and risks before approving any action.
Inspect the Agent Skill at https://github.com/camunda/camunda/blob/HEAD/.claude/skills/ci-validation/SKILL.md and all supporting files. Report provenance, permissions, scripts, dependencies, and security risks. Propose an exact installation into .agents/skills/ci-validation, but wait for my approval before writing files or running scripts.
Copying this prompt does not install or run the skill. Review third-party files before use.
CI Validation Skill
Use this skill for every workflow modification.
Mandatory Sequence
- Lint first:
actionlint .github/workflows/*.yml
- Policy check:
conftest test --rego-version v0 -o github --policy .github .github/workflows/*.yml
- Format:
./mvnw spotless:apply -T1C
- Tier classification (mandatory, before act):
- For each changed workflow, assign Tier 1/2/3 using
.claude/skills/act-testing/references/workflow-tiers.mdand the rubric in.claude/skills/act-testing/SKILL.md. - Record: workflow path, tier, one-line rationale, and validation path.
- For each changed workflow, assign Tier 1/2/3 using
- Act test (harness-style, mandatory):
- Read
.claude/skills/act-testing/SKILL.mdbefore proceeding. - Tier 1 or 2: create a temporary
test-*.ymlharness, runcheck-drift.sh, assessactfeasibility (act,docker, daemon), and provide user-run scenario commands. - Keep test harness files available while sharing commands so users can execute them.
- Remove
test-*.ymlfiles only after user-runactvalidation is complete. - Tier 3: skip harness and document non-applicability with explicit rationale.
- If multiple workflows changed, either test all Tier 1/2 workflows or explicitly justify why each one is Tier 3.
- Never run
actdirectly against production workflows.
- Read
- Iterate until checks are clean.
Do not skip step 1.
Required Evidence Format
For every changed workflow, include:
workflow: <path>tier: <1|2|3>reason: <single sentence>evidence: <drift-check + feasibility + scenario matrix | tier-3 rationale>
For Tier 1/2 workflows, evidence must explicitly contain:
test-workflow: <path>drift-check: <command + result>feasibility: act/docker/daemonscenario-matrix: <positive + negative at minimum>act-commands: <copy-paste ready plain text using canonical format: act <event> -e <event-json> -W <test-workflow> --secret-file .secrets --reuse; full workspace-relative paths; no markdown links in command text>cleanup: <test harness removed or justification>
If these fields are missing, validation is incomplete and should not be reported as done.
Validation Matrix
- Simple metadata/security updates:
actionlint + conftest + spotless - Logic-heavy workflows:
actionlint + conftest + spotless + act - External-only workflows:
actionlint + conftest + spotless + rationale
Preferred VS Code Tasks
CI: actionlint (workflows)CI: conftest (workflows policy)CI: verify skill markdown stability