seed: Skill for Claude Code

.agents/skills/run-release/SKILL.md

run-release is a skill for Claude Code, Codex from seed-hypermedia/seed. It costs 22 tokens per session (879 once invoked), scanned A, original, Apache-2.0.

A release runbook for preparing a production software release, including checks, tagging, continuous integration, release notes, and a latest.json file.

In plain words
What is it for?
Use it when releasing a new version from the main branch and following the project's documented release process.
Why use it?
It turns release work into a defined sequence and checks the code before creating and publishing a release tag.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one. Also seen: installed under .agents/ (shared by several agents).

This is seed-hypermedia/seed's own configuration. It tells Claude Code and Codex how to work on seed itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything seed configures →

Reuse

Borrowing it

Nothing to install: this file belongs to seed-hypermedia/seed. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/seed-hypermedia/seed/main/.agents/skills/run-release/SKILL.md
Clone the repo
git clone --depth 1 https://github.com/seed-hypermedia/seed

Made for: Claude Code, Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for run-release

README.md
[![agentmods](https://agentmods.dev/badge/skills/seed-hypermedia/seed/run-release/github.svg)](https://agentmods.dev/skills/seed-hypermedia/seed/run-release)
Your own site
<a href="https://agentmods.dev/skills/seed-hypermedia/seed/run-release"><img src="https://agentmods.dev/badge/skills/seed-hypermedia/seed/run-release/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for run-release

Your own site · 80×15
<a href="https://agentmods.dev/skills/seed-hypermedia/seed/run-release"><img src="https://agentmods.dev/badge/skills/seed-hypermedia/seed/run-release.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 22 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 879 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe. Third-party audits
  • NVIDIA SkillSpector pass 7 Sept 2026
How audits are shown
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 5.1 $0.00022 $0.00879
Opus 5.5 $0.00009 $0.00352
Sonnet 5.5 $0.00004 $0.00176
Haiku 4.5 $0.00002 $0.00088

Measured 20d ago against content hash fd7dc2cbabbb, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

run-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 20d 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.

.agents/skills/run-release/SKILL.md · 51 lines

How it starts

The opening of the file, as written. The whole thing — 51 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Run Release

Execute the release process in hypermedia/build/releasing.md. That doc is canonical for the process details; this skill is the interactive runbook. Invoking this skill counts as explicit permission to create and push the release tag, but confirm the version number with the user before pushing it.

Steps

  1. Preflight: confirm you are on main, the working tree is clean, and local main matches origin/main. Abort and report if not.
  2. Determine the version (YYYY.M.N, see hypermedia/build/releasing.md): check git tag --sort=-creatordate | head -5. If the latest tag is from the current month, increment its last number; otherwise start at .1 for the current month.
  3. Verify CI is green for the commit BEFORE tagging. The release workflows gate on the same frontend-tests (Lint, unit tests, e2e), so a red commit means a doomed release run and a re-tag. Check the existing result first — a push to main already triggered Dev - Docker Images on this exact commit: gh run list --workflow "Dev - Docker Images" --branch main --limit 5 --json headSha,status,conclusion,databaseId
    • Green for the commit → proceed.
    • Red → diagnose with gh run view <id>; fix on main first, then restart at step 1.
    • Still running or missing → don't wait for it; releases are often cut in a rush. Instead run the most obvious, fastest checks locally and then tag. In order of cost-effectiveness:
      • Root pnpm format:check (~1 min; the most common release-killer). The root script chains pnpm workspaces, then agents/, then vault/ with && — later groups are masked until earlier ones pass, so only the root script proves all of them.
      • Unit tests for packages touched since the last green CI run (e.g. cd frontend/apps/desktop && pnpm exec vitest run; cd agents && bun test).
      • Skip slow suites (e2e, backend, builds) — the release workflow runs full CI anyway; the local pass only exists to avoid tagging a commit that is dead on arrival.
  4. Confirm the version and the commit to be tagged with the user.
  5. Tag and push: git tag <version> && git push origin <version>.
  6. Wait for the release workflows (Release - Desktop App, Release - Docker Images) to go green: gh run list --workflow release-desktop.yml, then gh run watch <run-id>. Builds take tens of minutes — keep waiting, don't proceed early. If a workflow fails, stop and report.
  7. Draft release notes for <prev-tag>..<version> following the voice, grouping, and exclusion rules in the releasenotes skill, but match the exact format of the last few published releases (gh release view <prev-tag>). Key rules from hypermedia/build/releasing.md:
    • Very short feature entries; many commits often collapse into one line.
    • No entries for regressions introduced and fixed since the previous tag — never released, users never saw them.
    • End with the **Full Changelog** compare link.
  8. Show the draft to the user and apply their edits, then publish: gh release edit <version> --notes-file <tmpfile> --prerelease=false --latest (the workflow creates the release as a prerelease; this promotes it).
  9. Publish latest.json so desktop auto-update sees the release: gh workflow run "Generate latest.json (prod)" and confirm the run succeeds.
  10. Report: version, release URL, and the state of each step.

Read the full file on GitHub · 51 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. 20d ago Changed · -2 lines fd7dc2cbabbb
  2. 1mo ago First seen · 53 lines · 22 tokens per session scan A 87ca97b0e62c

Subscribe to this mod's changes

run-release is a skill published in the GitHub repository seed-hypermedia/seed (59 stars, last pushed today), licensed Apache-2.0. It adds 22 tokens to every session and 879 once invoked, about $0.0001 per session on Opus 5.5. 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.