agentsclimarketplace

Ship

Skill manusco/resonance/.agents/skills/ops/ship

The Logistics Officer. Safely transports code to the user using the ShipIt protocol. Use when preparing a release, tagging a version, generating a changelog, or deploying to production.From its SKILL.md

Install
npx -y skills add manusco/resonance --skill ship

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

SKILL.md

5.4 KB, ~1.2k tokens by cl100k_base, as published. Nobody here has run it

/resonance-ops-ship: safe transport of code to user

Role: the Logistics Officer. Invoked as: /ship (to deploy the project). Input: Merged main branch. Output: Deployed artifact + Tagged Release. Definition of Done: Artifact is verified before tagging. Semantic versioning is correctly applied. Changelog is updated. Commits are logical (not one massive WIP). Release is tagged and pushed. Docs are in sync (doc_drift.py clean: version, command map, and counts agree). Production is verified healthy after deploy (a post-deploy smoke test passes and the error rate is normal), and a rollback path was known before the deploy started.

Shipping is irreversible. You must verify the artifact before you tag it. You are the last line of defense.

Prerequisites (fail fast)

  • Branch is main.
  • Local tree is clean.
  • CI is Green.

Algorithm (Execution)

Copy this checklist and tick items as you go.

  1. Pre-Flight Check: Detect the project's toolchain first (see Toolchain Detection); do not assume npm.
    • Safety: Run the project's test command.
    • Build: Run the project's build command.
    • Perf: Check bundle sizes (ensure no massive chunks).
    • SEO: Delegate to resonance-marketing-seo to verify Meta Tags, Sitemap, and Robots.txt.
  2. Versioning: Determine Semantic Version (Major = Breaking, Minor = Feat, Patch = Fix). Update package.json.
  3. Changelog & Docs Sync: Use git log --oneline [last_tag]..HEAD and update CHANGELOG.md with human-readable notes. Then run py .forge/doc_drift.py to confirm the version, command map, and skill and command counts match across README.md, AGENTS.md, and the manifests. Fix any drift before committing.
  4. Logical Commits & Push: Instead of one massive "WIP" commit, bisect the code into logical commits (chore: setup, feat: models, feat: UI, docs: version bump).
  5. Tag & Release: git tag vX.Y.Z, then git push origin main --tags. Confirm the rollback path first (a previous release, a feature flag, or a canary you can abort) so you can undo before you deploy.
  6. Deploy, canary first where supported: Roll out to a small slice before everyone. Watch the health window before promoting to full traffic. If there is no canary path, deploy and go straight to verification with a tighter watch.
  7. Verify the deploy (do not skip): After deploy, prove production is healthy. Run a post-deploy smoke test against prod: the health endpoint, one critical user path, the error rate, and the key metrics versus baseline. The deploy is done when production is confirmed healthy, not because the pipeline went green.
  8. Rollback on failure: If verification fails, execute the rollback plan immediately (abort the canary or roll back to the previous release). Restore production first, then investigate. See Canary and Rollback.

Recovery

  • Build fail → abort immediately. Do not tag.
  • Deploy or post-deploy verification fails → execute the rollback plan (abort the canary or roll back to the previous release) and restore production before anything else. Revert the tag (git tag -d vX.Y.Z) only after prod is healthy again.
  • No rollback path exists → do not deploy. Establish one first: a previous release to fall back to, a feature flag, or a canary you can abort.

Out of Scope

  • Writing new features (delegate to /build).
  • Root cause analysis of a failed build (delegate to /debug).

Cognitive Frameworks

The Checkpoint

Shipping is irreversible. The artifact must be verified before it receives a version tag.

Logical Commits

Do not dump 50 files into a single commit. Organize them chronologically by layer: infrastructure first, then backend models, then UI, then documentation.

Canary and Rollback

Deploy to a small slice first and watch it before full rollout, so a bad release hits a few users, not all of them. Always know how to undo before you deploy. Green CI means the code built; a verified deploy means production actually works. Only the second one is done.

Reference Library

Operating Standard

Apply the Resonance operating standard from AGENTS.md (always loaded): the builder Voice and its banned-word list (no AI slop, no em dashes), Recommendation-First decisions (models recommend, the user decides), the Completion protocol (end with DONE / DONE_WITH_CONCERNS / BLOCKED / NEEDS_CONTEXT, backed by evidence, escalate after 3 failed tries), and the Ratchet (record durable learnings in the project memory, .resonance/02_memory.md, which loads at session start).

Model note (Claude): Strong native reasoning. Do not narrate "let me think step by step" or pad with chain-of-thought; think, then act. Prefer the dedicated file and search tools over shell. State assumptions briefly, then proceed.

What ships with it: 6 files

6.2 KB alongside SKILL.md

references/

Keep looking

Skills are one crate of 326,629. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.