agentleFS
ecorkranecorkran/context-forge/.agents/skills/release/SKILL.md

release

Skill5 starsChanged 7 days agoMITView on GitHub

Cut a context-forge release — confirm npm login, verify cf state, merge the slice, build, bump all four packages in lockstep, update CHANGELOG/DEVLOG, tag, push, publish to npm, and confirm the new version is live. Use when the user says "release", "cut a release", "ship it", or "bump and tag".

  • Deletes or force-pushes
  • Commits and pushes

196 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use create-agent-adapter, changelog-style and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • MicrosoftSkill

    create-agent-adapter

    Technical guide for creating a new Paperclip agent adapter. Use when building a new adapter package, adding support for a new AI coding tool (e.g. a new CLI agent, API-based agent, or custom process), or when modifying the adapter system. Covers the required interfaces, module structure, registration points, and conventions derived from the existing claude-local and codex-local adapters.

    Also in 5 projects that use this

  • GrinvSkill

    changelog-style

    Turn a commit range into CHANGELOG.md entries following common-changelog style. Use when finishing a PR or cutting a release, to write/fix the [Unreleased] section's entries — moving [Unreleased] under a new dated version heading is a separate, automated step (see the release skill / scripts/sync-version.ts), not this skill's job.

    Also in 4 projects that use this

  • GrinvSkill

    docs-consistency-check

    Check README/manifest.json/server.json/CHANGELOG.md/AGENTS.md and docs/*.md for drift against the actual registered tools and source. Use after adding, renaming, or removing a tool, or as part of a live-audit pass.

    Also in 4 projects that use this

  • Microsoft AzureSkill

    component-release

    Prepares a verifiable Agent Landing Zone UI release. Use for release branches, VERSION, changelog entries, tags, release notes, and compatibility evidence; never publish without explicit human approval.

    Also in 4 projects that use this

  • Microsoft AzureSkill

    documentation-consistency

    Keeps AI Landing Zone contributor, user, and operator documentation aligned with shipped Bicep behavior. Use for parameters, defaults, outputs, topology, deployment flow, operations, or releases.

    Also in 8 projects that use this

  • rivet-devSkill

    launch-graphics

    Generate agentOS marketing/launch graphics (launch & social heroes, code-snippet cards) from the committed generators in docs-internal/graphics. Use when asked to make, update, or render an agentOS launch image, social image, "agentOS Apps" graphic, or technical code-snippet image.

    Also in 3 projects that use this

  • +3 more from Grinv

The file

We split this skill into the rules it gives an agent

12 rules. 0 are common, 3 are distinctive, and 9 are very specific to this project.

  • 0

    Common

    Most of the 186 edited versions we read keep these.

  • 3

    Distinctive

    This author's own choices. Worth a look.

  • 9

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Run from the repo root. Stop and report on any failed command.

    line 8
  2. Very specific:

    Run npm whoami. If it fails (401 / not logged in), stop and ask the PM to run npm login.

    line 12
  3. Very specific:

    Run cf next. If it reports something unfinished for the current slice (open review gate, unchecked tasks, failing check), stop and show it.

    line 16
  4. Very specific:

    Read the target with cf config get git.integration_branch (empty means main).

    line 20
  5. Very specific:

    Run pnpm -r build. A failure stops the release.

    line 26
  6. Very specific:

    Read ## [Unreleased] in CHANGELOG.md.

    line 32
  7. Distinctive:

    Minor: only for a new user-facing feature or a significant new AI capability.

    line 35
  8. Distinctive:

    Major: only when the PM explicitly asks for it.

    line 36
  9. Very specific:

    Run pnpm -r typecheck and pnpm -r test.

    line 52
  10. Very specific:

    Run pnpm -r build again. dist/ is gitignored and packed straight from disk, so this must run right before publish — a stale dist has …

    line 67
  11. Very specific:

    Run pnpm publish -r --access public.

    line 68
  12. Very specific:

    Never check context-forge unscoped — that is an unrelated third-party package.

    line 85

“Common” is measured: kept by at least 6 in 10 of the 186 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Release context-forge

Run from the repo root. Stop and report on any failed command; don't work around it.

0. PM prerequisite: npm login

Run npm whoami. If it fails (401 / not logged in), stop and ask the PM to run npm login. Do not continue until npm whoami succeeds.

1. Check cf state

Run cf next. If it reports something unfinished for the current slice (open review gate, unchecked tasks, failing check), stop and show it. Otherwise continue.

2. Merge

Read the target with cf config get git.integration_branch (empty means main). If the current branch is a slice branch not yet merged into the target, follow the merge steps in AGENTS.md (Git Rules). If already on the target with the work merged, skip this.

Working tree must be clean before continuing.

3. Build

Run pnpm -r build. A failure stops the release.

4. Pick the version

The four publishable packages release in lockstep at one version: packages/core, packages/cli, packages/mcp-server, packages/context-forge. Current version is version in packages/core/package.json. The repo-root package.json is private at 0.0.0 — never touch it.

Read ## [Unreleased] in CHANGELOG.md. If it is empty, draft entries from git log vPREV..HEAD --oneline (user-facing feat/fix commits, with issue numbers) and add them before continuing.

  • Patch is the default. It covers fixes, refinements, and new flags, options, parameters, config keys and checks. An ### Added section or a non-empty changelog does not by itself mean minor.
  • Minor only for a new user-facing feature or a significant new AI capability. Propose it in one line with the reason, and wait for the PM to confirm before changing any files.
  • Major only when the PM explicitly asks for it.

For a patch, state the chosen version in one line and proceed.

5. Bump and document

  1. Set version to X.Y.Z in all four package.json files listed above. Internal deps are workspace:* and need no edit.
  2. CHANGELOG.md: insert ## [X.Y.Z] - YYYYMMDD directly under ## [Unreleased], so the unreleased entries now sit under the new version and [Unreleased] is empty. Entries stay short and user-facing.
  3. DEVLOG.md: under today's ## YYYY-MM-DD heading (create it at the top, below the ---, if missing), add at the top:
    ### Release X.Y.Z
    
    - **Contents:** <slices and fixes included, with issue numbers>
    
    Tags: @context-forge/core@X.Y.Z, @context-forge/cli@X.Y.Z, @context-forge/mcp@X.Y.Z, @context-forge/context-forge@X.Y.Z
  4. Run pnpm -r typecheck and pnpm -r test. Any failure stops the release.
  5. Commit: chore: release X.Y.Z.

6. Tag and push

git tag -a vX.Y.Z -m "vX.Y.Z"
git push
git push --tags

The tag must point at HEAD, or pnpm's git checks block the publish. If a fix lands after tagging, move it with git tag -f -a vX.Y.Z -m "vX.Y.Z" and git push -f origin vX.Y.Z.

7. Publish

  1. Run pnpm -r build again. dist/ is gitignored and packed straight from disk, so this must run right before publish — a stale dist has shipped before.
  2. Run pnpm publish -r --access public. The PM approves each package's publish as npm asks (2FA). If it fails with EOTP, nothing was published: ask the PM for an OTP and re-run with --otp=<code>, or hand the command to the PM to run.

8. Confirm on npm

Packages publish one at a time, each gated by its own PM approval, so a full run can take several minutes. Don't check until the PM says the publish finished.

Query the registry directly — npm view goes through a cache that lags behind a fresh publish:

for p in core cli mcp context-forge; do
  echo "$p: $(curl -s "https://registry.npmjs.org/@context-forge%2F$p?t=$(date +%s)" \
    | python3 -c 'import sys,json;d=json.load(sys.stdin);print(d["dist-tags"]["latest"], d["time"].get("X.Y.Z","MISSING"))')"
done

Each package must show X.Y.Z as latest with a publish timestamp. If any is missing, re-check every 30 seconds for up to 5 minutes before calling it a failure. Only a package still missing after that is a real partial publish: re-run pnpm publish -r --access public (it skips versions already published).

Never check context-forge unscoped — that is an unrelated third-party package.

Report "X.Y.Z live on npm" when all four match.

9. Summary

Invoke /sq:summary (Skill tool, skill sq:summary) as the last action of the release. Call it even when the release is paused at step 7 waiting for the PM to run pnpm publish -r --access public (or supply an OTP): hand over the command, call /sq:summary right then, and don't wait for the publish to finish. Step 8 happens after the PM reports back, in the same session or a restored one.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.