ship

A deployment step for the Compass development process. It takes a finished build to production, checks it against production data, and verifies that its monitoring signal is emitted when deployment is in scope.

In plain words
What is it for?
Use it after a build is closed to deploy through the repository's normal process, reconcile production data, and check observability.
Why use it?
It closes the gap between a build passing local or staging checks and proving that the deployed version works in production.

Skill for Claude CodeCodex

Part of the compass plugin — 11 skills, 3 commands, 3 hooks shipped together

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/rishi4792/compass/ship
Any agent
npx skills add Rishi4792/compass --skill ship
Clone the repo
git clone --depth 1 https://github.com/Rishi4792/compass

Made for: Claude Code, Codex.

Or install compass, the plugin that ships this one along with the rest of its 11 skills, 3 commands, 3 hooks.

Per session 72 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 7,388 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00072 $0.07388
Opus 5 $0.00036 $0.03694
Sonnet 5 $0.00014 $0.01478
Haiku 4.5 $0.00007 $0.00739

Measured 2d ago against content hash 0f61a02f1c1a, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

ship scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

plugins/compass/skills/ship/SKILL.md · 192 lines

How it starts

The opening of the file, as written. The whole thing — 192 lines — stays where its author put it; the contents beside it link to each section on GitHub.

compass:ship

The lifecycle verifies locally (prod stays read-only during build). This stage takes a CLOSED build to production and proves it there — closing the gap where the contract's Observability check only means something post-deploy.

Read these before you start. They are the standards this stage is held to, and they live in plugins/compass/shared/ so they are the same for every stage that uses them:

  • shared/feynman.md — the writing standard for the reader-facing copy this stage produces.

(A standard nobody loads is not a standard. shared/MANIFEST declares who reads each file and doctrine-wired-check.sh proves it — feynman.md sat unread for three releases while its own first line claimed three stages loaded it.)

When NOT to run

If the contract's Non-goals mark deploy out of scope, skip — the build is done at CLOSED and the observability check was scoped to staging. Say so and stop.

Step 0 — own, claim the ship lock (single-flight), then gate

  1. Own this build: compass.sh own <slug> --session "$CLAUDE_CODE_SESSION_ID" (the Stop hook guards this session through ship).
  2. Claim the ship lock FIRST and unconditionally (v0.9.0 single-flight): compass.sh ship-claim <slug>. Non-zero → STOP — another build holds the lock (it names the holder); only one build per project ships at a time. The lock self-heals (steals a SHIPPED/ROLLED-BACK or >2h-stale holder), so a crashed ship never deadlocks future ships. You MUST compass.sh ship-release <slug> on EVERY exit from ship — success (SHIPPED), yield (Step 0.4), or any hard-stop (prod unreachable) — so the lock is never leaked.
  3. Gate: compass.sh gate .claude/builds/<slug> review-build. Non-zero → STOP (build not CLOSED/signed-off; ship-release first), offer compass:review-build.
    • The one exception, and it is the user's to make (v0.30): if the review-build receipt reads ACCEPTED WITH OPEN FINDINGS, the gate refuses — correctly. A review that did not converge may still ship, but ONLY when the user has signed for it. Run compass.sh converge-waiver .claude/builds/<slug>; it exits 0 only when the receipt carries a - [x] converge-waiver: user-signed · <what is open, and who accepted it> line, and it prints every unmet item to stderr every time. Non-zero → STOP. A model-authored header does not count — that is how cold-critic became switchable. This check deliberately does NOT live inside compass.sh gate: v0.28's INV-NO-LIFECYCLE-CHANGE freezes that function's PASS/SUPERSEDED/unchecked-box semantics byte-for-byte, and a lifecycle change must never be made in passing.
    • When the waiver path is used, everything downstream must say so. The Release Card carries shipped un-converged — <one line naming what is open>, and the ship receipt records converge-waiver: honoured. A build that ships with known open findings and does not say so on its own release page is the defect this rule exists to prevent. Read contract.md (deploy/rollback/observability are the invariant here).

Read the full file on GitHub · 192 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 192 lines · 72 tokens per session scan A 0f61a02f1c1a

Subscribe to this mod's changes

ship is a skill published in the GitHub repository Rishi4792/compass (2 stars, last pushed 8d ago), licensed MIT. It adds 72 tokens to every session and 7,388 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other skills, from other repositories

kagent-dev

Development guide for kagent's v1alpha3 Harness and AgentTemplate CRDs, AgentInstance gRPC control plane, upstream A2A integration, Substrate runtime provisioning, tests, generation, and PR workflow. Use for any implementation, debugging, review, or CI task in the kagent repository.

kagent-dev/kagent · 65 tokens

timoni

Use when deploying applications to Kubernetes with Timoni. Covers installing and upgrading module instances from OCI registries, composing multi-app deployments with bundles, injecting values from clusters or CI with runtimes, targeting multiple clusters, and authoring, testing, signing and publishing modules with CUE.

stefanprodan/timoni · 59 tokens

azmon-mirroredcatalogs-operations-cli

Brings Azure Monitor, Application Insights, and Log Analytics telemetry into Fabric as Eventhouse external delta tables and correlates it with business data. Use to onboard observability data, judge whether latency or availability affected revenue, or build a Real-Time dashboard and Operations Agent over it.

microsoft/skills-for-fabric · 66 tokens

durable-objects

Create and review Cloudflare Durable Objects. Use when building stateful coordination (chat rooms, multiplayer games, booking systems), implementing RPC methods, SQLite storage, alarms, WebSockets, or reviewing DO code for best practices. Covers Workers integration, wrangler config, and testing with Vitest. Biases…

cloudflare/skills · 76 tokens

ship-web-games

Package, deploy, and verify a playable Three.js or web game. Use for release builds, asset delivery, private/public deployment, production smoke tests, browser proof, release notes, rollback readiness, and cleanup of temporary QA resources.

MengTo/Skills · 50 tokens

hyperpod-version-checker

Check and compare software component versions on SageMaker HyperPod cluster nodes - NVIDIA drivers, CUDA toolkit, cuDNN, NCCL, EFA, AWS OFI NCCL, GDRCopy, MPI, Neuron SDK (Trainium/Inferentia), Python, and PyTorch. Use when checking component versions, verifying CUDA/driver compatibility, detecting version mismatches…

awslabs/agent-plugins · 120 tokens