release

A release workflow that moves approved code from a release branch into the production branch and handles the related review, deployment, documentation, and optional release notes.

In plain words
What is it for?
Use it to prepare a release pull request, run the required review, merge after approval, update documentation, and optionally create release notes.
Why use it?
It organizes the steps around a production release and keeps the final merge under explicit owner approval.

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/devstride/claude-plugin/release
Any agent
npx skills add devstride/claude-plugin --skill release
Clone the repo
git clone --depth 1 https://github.com/devstride/claude-plugin

Made for: Claude Code, Codex.

Per session 72 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 5,572 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 $0.05572
Opus 5 $0.00036 $0.02786
Sonnet 5 $0.00014 $0.01114
Haiku 4.5 $0.00007 $0.00557

Measured yesterday against content hash 9f5925451d5e, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

release 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 yesterday.

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.

skills/release/SKILL.md · 237 lines

How it starts

The opening of the file, as written. The whole thing — 237 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Human output. Read ${CLAUDE_PLUGIN_ROOT}/skills/build-item/references/plain-language-output.md once per top-level run; composed skills reuse it. Apply it to every message.

Promote release.releaseSourcerelease.productionBranch (default developmaster). That merge triggers release.autoDeployOnMerge, so prepare autonomously but require explicit owner approval before merging. After confirmed deploy, invoke the repo's local docs.updateSkill unless no docs; write release notes only when --release-notes explicitly asks.

Optional arguments — documentation switches and a scope: $ARGUMENTS

  • --release-notes <false|true|draft> — absent = false. Bare flag/“release notes” = true; draft leaves it unpublished. Step 5c runs only after confirmed merge and live deploy.
  • no docs / skip docs — suppress the core-documentation update (step 5b) for this run. Docs are otherwise updated by default whenever a docs skill is registered.
  • docs only — run step 5b alone against an already-shipped release, with no PR and no merge. release-notes only — run step 5c alone against an already-shipped release, still subject to the deploy confirmation. Asking for this scope IS the request: it implies --release-notes true, or draft when the word draft accompanies it (release-notes only draft).

Repo config. Load .claude/ds-config.json first and treat it as authoritative: the keys this skill reads are release.productionBranch, release.releaseSource, release.autoDeployOnMerge (a plain-English string quoted back to the owner — never assume a provider), release.deployVerification, docs.updateSkill and docs.releaseNotesSkill (LOCAL skill names; their contract is ${CLAUDE_PLUGIN_ROOT}/skills/release/references/docs-hooks.md, the authority over anything restated here), plus baseBranch / protectedBranches / ci.freezeBaseWhileReleasePrReady and the review.* / verify.* / preShipChecks / prBodyTemplate blocks the composed skills consume. Inline literals are shipped defaults — if the file disagrees, the file wins; if it is absent, fall back to them and say so.

IMPORTANT — autonomy boundary:

  • Run steps 0–3 (delta → PR → full gated review → documentation preparation) autonomously; only surface genuine forks (an ambiguous/risky/unverifiable review finding, an unresolvable conflict, or an infra/secret gate that is the owner's to provision).
  • The master merge is a hard gate. NEVER merge developmaster without an explicit, in-the-moment owner go-ahead — it deploys production. Present the release summary and ask. Invoking /devstride:release is intent to PREPARE a release, not standing authorization to deploy.
  • The core-documentation update is pre-authorized by default whenever docs.updateSkill is registered — the local skill decides what publishing means and may pause for a look — but it is still an outward action: explain the result plainly and preserve exact pages, links, branch and pull request. It runs only after the merge is confirmed (step 5b); nothing is published before the owner approves the production merge. "no docs" skips it.
  • Release notes are NEVER written without --release-notes. There is no size, scope or "user-facing" threshold this skill interprets on the owner's behalf; the flag is the only trigger, and the note comes after the confirmed deploy (step 5c). If it seems a release deserves a note the owner did not ask for, say so in the step-4 summary and let them add the flag.

IMPORTANT — the DevStride MCP targets PRODUCTION; git/gh act on the real repos. Treat the production merge, and anything the local docs skills publish, as real, user-visible production events.

0. Preconditions and the release delta

  • Confirm this is a release. The release is release.releaseSource (develop) → release.productionBranch (master). If invoked standalone with the tree on some other branch, that's fine — this skill operates on the two named branches, not the current checkout. If a develop → master PR is already open, adopt it instead of cutting a duplicate.
  • Prove CI can stay last before opening anything. Inspect pull-request workflows, excluding the convention-only shape in ${CLAUDE_PLUGIN_ROOT}/skills/setup/references/ci-cost-patterns.md. None → valid no-CI release, said plainly. Any expensive PR workflow without all draft-hold flags true → STOP with /devstride:setup ci; false, mixed or ambiguous facts cannot prove a hold.
  • Sync both branches. git fetch origin master develop. Confirm origin/develop is ahead of origin/master (there is something to release); if not, STOP and say there's nothing to release.
  • Settle the release source BEFORE cutting (when ci.freezeBaseWhileReleasePrReady is true, the default; false skips this and the freeze, and says so): list every open NON-DRAFT PR into releaseSource — green, pending or red alike — and decide each NOW: merge it first (it joins this release; the delta is recomputed) or park it (gh pr ready --undo) until this release merges. Say which choice was made for each and keep the parked list — step 6 un-parks them.
  • Record <sourceHead> — the SHA of origin/<releaseSource> the delta was computed from — whatever the freeze switch says. Every later integrity check compares against this immutable value, never against the branch tip (the release PR's head IS the branch, so the tip and "the PR head" move together and comparing them proves nothing).
  • Compute the delta — what this release ships. Read git log --first-parent origin/master..origin/develop and the merged PRs in that range (gh pr list --base develop --state merged cross-referenced to the range) to assemble: the epics/stories that landed, and — importantly — which changes are user-facing (new/changed UI, API, behavior, permissions, migrations) vs internal. This delta drives BOTH the PR body and the docs pass. Note anything that looks like a breaking change or a migration explicitly.
  • Report the delta summary (epic/story list + the user-facing subset) before cutting the PR.

Read the full file on GitHub · 237 lines

Files

What ships with it

2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

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. yesterday First seen · 237 lines · 72 tokens per session scan A 9f5925451d5e

Subscribe to this mod's changes

release is a skill published in the GitHub repository devstride/claude-plugin (2 stars, last pushed 3d ago), licensed MIT. It adds 72 tokens to every session and 5,572 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

systematic-debugging

Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.

obra/superpowers · 21 tokens

next-cache-components-adoption

Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…

vercel/next.js · 95 tokens

babysit-pr

Babysit a GitHub pull request after creation by continuously polling review comments, CI checks/workflow runs, and mergeability state until the PR is merged/closed or user help is required. Diagnose failures, retry likely flaky failures up to 3 times, auto-fix/push branch-related issues when appropriate, and keep…

openai/codex · 114 tokens

imagegen

Generate or edit raster images when the task benefits from AI-created bitmap visuals such as photos, illustrations, textures, sprites, mockups, or transparent-background cutouts. Use when Codex should create a brand-new image, transform an existing image, or derive visual variants from references, and the output…

openai/codex · 113 tokens

cpu-profile-analysis

Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…

microsoft/vscode · 71 tokens

next-cache-components-optimizer

Drive a Next.js route to instant navigation by setting up an agentic loop, under Cache Components / PPR, on initial load (hard navigation) and client-side navigation (soft navigation). Encode the goal as a failing @next/playwright instant() e2e and work it to green, one verified route at a time; the shipped test then…

vercel/next.js · 170 tokens