Keep an app repo current with Eve conventions. The checks are independent — run them in parallel when multiple areas need attention.
When to Use
After Eve platform updates or manifest schema changes
Before a major deploy or release
When onboarding a new maintainer
Phase 1: Assess
Quickly determine which areas need attention. Check which of these files exist and whether they look stale — don't deep-read every file, just note which areas need work:
.eve/manifest.yaml — does it exist? Is the schema line eve/compose/v1?
skills.txt — does it exist? Are there pinned or obsolete entries?
AGENTS.md / CLAUDE.md — do they reference current skills?
Codebase — any obvious deprecated CLI patterns or inline secrets?
If only one area needs work, handle it directly. Otherwise, proceed to Phase 2.
Phase 2: Dispatch Workers
Create one worker per area that needs updating. Each worker description below is self-contained — a worker can execute its area independently without context from the others.
Worker: Manifest Alignment
Check and fix .eve/manifest.yaml:
Ensure schema: eve/compose/v1 is present.
Prefer services: over legacy components:.
Keep x-eve ingress and pipeline definitions accurate.
Keep x-eve.defaults in sync with harness defaults (harness/profile/options).
Keep x-eve.agents profiles aligned with orchestration policy.
Ensure x-eve.agents.config_path and x-eve.chat.config_path point to valid files.
Confirm ${secret.KEY} usage for secrets.
Deploy pipelines should include a build step before release.
Services with Docker images should have build.context defined.
Registry auth secrets (REGISTRY_USERNAME + REGISTRY_PASSWORD) are required only for custom BYO registries.
Worker: Skills File
Check and fix skills.txt:
Keep Eve skillpack references up to date.
Remove obsolete packs or pinned versions.
Worker: Agent Instructions
Check and fix AGENTS.md, CLAUDE.md, and agents/ config files:
Update skill references to include eve-se-index.
Remove stale commands or URLs.
agents/agents.yaml defines agents and skills — verify entries are current.
agents/teams.yaml defines team composition and dispatch — verify structure.
agents/chat.yaml defines chat routing rules and permissions — verify rules.
Worker: Deprecated Patterns
Search the codebase for deprecated patterns and fix or flag them:
Old CLI commands (eve deploy vs eve env deploy)
Old deploy syntax without --ref parameter
Hardcoded domains in docs or manifests
Inline secrets in repo files
Dockerfiles missing org.opencontainers.image.source label pointing to the repo URL
Pipelines missing build step before release
Services with Docker images but no build.context configuration
Missing registry authentication secrets for custom registries (REGISTRY_USERNAME, REGISTRY_PASSWORD or provider-equivalent names)
Phase 3: Verify
After all workers complete, run final verification:
# Local validation (Docker Compose)
docker compose up --build
# Staging deploy (requires --ref with 40-char SHA or a ref resolved against --repo-dir)
eve env deploy staging --ref main --repo-dir .
# Use --direct to bypass pipeline if needed
eve env deploy staging --ref main --repo-dir . --direct
