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.
npx agentmods add commands/joncik91/ucai/shipgit clone --depth 1 https://github.com/Joncik91/ucaiWhat 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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00011 | $0.04798 |
| Opus 5 | $0.00005 | $0.02399 |
| Sonnet 5 | $0.00002 | $0.00960 |
| Haiku 4.5 | $0.00001 | $0.00480 |
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.
How it starts
The opening of the file, as written. The whole thing — 354 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Autonomous Ship Pipeline
You are an autonomous development agent. Your job is to take a spec and deliver a working PR — no human intervention needed between start and finish.
This is NOT /build. There are ZERO approval gates. You make every decision yourself: architecture, implementation, testing, PR. The human reviews the PR, not the intermediate steps.
Core Principles
- Autonomous execution: Never stop to ask the user. Make the best decision and move forward.
- Verify everything: Run tests and linters deterministically. If they fail, fix and retry.
- Commit per milestone: Each milestone gets its own commit for clean PR history.
- Fail gracefully: If you hit max fix attempts, proceed with warnings — don't block the whole pipeline.
- Track progress: Update ship state file phase after each phase completes.
- Engine enforcement: A ContingencyEngine tracks dependencies and gates. Before each phase, check gates. If a gate blocks, auto-remedy (complete the prerequisite) or degrade to a warning — never stop to ask the user. A degrade-to-warning decision must be recorded explicitly in the run output (ship state file and/or PR description) — never applied silently. After each phase, update engine state.
Skill Loading
Before Phase 2, identify and load relevant skills automatically:
- Determine work type from the spec (backend, frontend, full-stack, infrastructure)
- Load matching skill(s):
Skill(ucai:backend),Skill(ucai:frontend),Skill(ucai:architect),Skill(ucai:qa),Skill(ucai:devops) - Apply skill guidance throughout all phases
Phase 0: Setup
Goal: Parse arguments, set up isolation, load context.
Actions:
- Initialize the ship pipeline:
echo '$ARGUMENTS' | Bash(node "${CLAUDE_PLUGIN_ROOT}/scripts/setup-ship.js" --stdin) - Read the created state file
.claude/ucai-ship.local.mdto confirm setup - Unless
--no-worktreewas specified, enter a worktree for isolation:- Generate a branch name from the spec:
ship/<slug>-<timestamp> - Use the worktree to isolate all changes from the user's working directory
- Generate a branch name from the spec:
- Load project context (skip silently if files don't exist):
.claude/project.md— project vision, tech stack, constraints.claude/requirements.md— backlog, build orderCLAUDE.md— project conventionstasks/lessons.md— known patterns and past corrections
- Initialize engine:
Bash(node "${CLAUDE_PLUGIN_ROOT}/scripts/setup-ship-engine.js" --spec "$ARGUMENTS") - Update engine:
Bash(node "${CLAUDE_PLUGIN_ROOT}/scripts/update-engine.js" --pipeline ship --dep dep-ship-state-init --state complete --proof "state file created") - Update ship state: set
phase: 0complete
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.
- 2d ago First seen · 354 lines · 11 tokens per session scan A 401f02237e81
ship is a command published in the GitHub repository Joncik91/ucai (29 stars, last pushed 7d ago), licensed MIT. It adds 11 tokens to every session and 4,798 once invoked, about $0.0001 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 2026-08-30.
Other commands, from other repositories
quick-commit
Create a conventional commit with an auto-generated message.
commit
Create well-formatted commits with conventional commit messages and emojis.
commit
Analyze git diffs or staged changes and generate conventional commit messages that explain WHY a change was made. Supports auto-detecting type and scope, intelligent file staging, and interactive overrides. Use when asked to "write a commit message", "generate a commit", "describe my changes", "commit this"…
fest-commit
Commit changes with festival traceability metadata.
feature-flow-w
Complete git worktree workflow orchestration (GitHub/GitLab/Forgejo) - worktree, branch, issue, commit.
feature-flow
Complete git workflow orchestration (GitHub/GitLab/Forgejo) - branch, issue, commit.