ship-commits

A release workflow for sending existing local Git commits through a project’s merge queue, which automatically checks and merges approved pull requests. It uses a separate worktree, then creates, pushes, and queues the pull request.

In plain words
What is it for?
Use it when one or more finished commits need to be sent to `main` through the repository’s pull-request and merge-queue process.
Why use it?
It reduces the risk of including unrelated work or disturbing other changes in the main working directory.

Skill for Claude CodeCodex

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/phase-rs/phase/ship-commits
Any agent
npx skills add phase-rs/phase --skill ship-commits
Clone the repo
git clone --depth 1 https://github.com/phase-rs/phase

Made for: Claude Code, Codex.

Per session 115 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 4,025 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 $0.00115 $0.04025
Opus 5 $0.00057 $0.02013
Sonnet 5 $0.00023 $0.00805
Haiku 4.5 $0.00012 $0.00402

Measured 3d ago against content hash 0858189f42bd, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

ship-commits 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 3d 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.

.claude/skills/ship-commits/SKILL.md · 285 lines

How it starts

The opening of the file, as written. The whole thing — 285 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Ship Commits

Take local commits and land them on main through the merge queue, using an isolated worktree to avoid carrying other agents' concurrent work into the PR.

Why a worktree

The main working dir is assumed to contain other unrelated changes from concurrent agents — unstaged work, other branches, in-progress commits on main that aren't yet shipped. Branching in place risks:

  • Carrying unrelated commits into the PR (if local main is ahead of origin/main with multiple agents' work).
  • Disturbing other agents' branch state with git checkout operations.
  • Leaving the user on a topic branch they didn't expect to be on.

A worktree based on origin/main gives a clean branch we cherry-pick into, isolated from everything else.

When to use

  • User has finished a discrete chunk of work and wants it shipped.
  • One or more commits exist locally (named explicitly, or "the last N commits on main", or "this commit").
  • Goal: branch → push → PR → enqueue, with the queue handling rebase + CI + merge in the background.

When NOT to use

  • Work is uncommitted. Run /commit first (commit by pathspec — your memory's feedback_shared_index_commit_pathspec). Uncommitted work being shipped must not be left dirty on local main.
  • A PR for these commits already exists. Use gh pr merge <N> --auto directly to enqueue.
  • The commits are already on origin/main. Nothing to do.
  • Repo isn't phase-rs/phase. This skill encodes phase.rs-specific conventions.

Phase.rs-specific conventions

Convention Why
The merge queue owns the merge strategy — do NOT pass --squash. gh pr merge --squash --auto is rejected with The merge strategy for main is set by the merge queue. Pass --auto alone. The queue still produces a squash commit, which is why Step 0's reconciliation must stay squash-aware. Repo keeps main as one-commit-per-feature, enforced queue-side rather than client-side.
--auto is mandatory on gh pr merge. Without it, the merge bypasses the queue and tries to merge immediately — fails on protected main.
--no-verify on push. Pre-push hooks duplicate validation that Tilt/CI has already done locally. CI/queue will re-validate.
Branch protection on main blocks direct push for non-admins. Queue handles serialization for parallel PRs.
Worktree-based shipping. Main working dir has concurrent changes from other agents — don't disturb them.

Read the full file on GitHub · 285 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. 3d ago First seen · 285 lines · 115 tokens per session scan A 0858189f42bd

Subscribe to this mod's changes

ship-commits is a skill published in the GitHub repository phase-rs/phase (259 stars, last pushed 3d ago), licensed Apache-2.0. It adds 115 tokens to every session and 4,025 once invoked, about $0.0006 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.