agentleFS
NVIDIAKnown publisherNVIDIA/NeMo-Relay/.agents/skills/prepare-code-freeze/SKILL.md

prepare-code-freeze

Skill238 starsChanged 2 months agoApache-2.0View on GitHub

Execute a requested NeMo Relay code freeze by cutting a release branch, updating nightly alpha configuration, bumping main, and preparing the required PR. Do not use for an ordinary version bump or release-note draft.

  • Commits and pushes

1 other project uses a version of this, 1 of them edited.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

10 rules. About 0 are common, 5 are distinctive, and 5 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈5

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 5

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use this skill when the user asks to start, prepare, or automate a NeMo Relay code freeze.

    line 9
  2. Distinctive:

    If the remote release branch already exists, verify it points where expected before continuing.

    line 29
  3. Very specific:

    Create a PR branch from latest upstream/main, for example chore/version-bump-<major>.<minor>.

    line 31
  4. Very specific:

    Update .github/nightly-alpha-branches.yaml to include the new release branch.

    line 33
  5. Very specific:

    Run just set-version <next-version> to update all project-owned unified-release version surfaces on main, including Cargo, Python, Node, …

    line 35
  6. Very specific:

    Search documentation source for references to the old version and update current-version install commands, package examples, and …

    line 44
  7. Very specific:

    After explicit approval to publish, open a PR targeting main using .github/pull_request_template.md.

    line 68
  8. Distinctive:

    Do not create release tags.

    line 79
  9. Distinctive:

    Do not target the code-freeze PR at the release branch.

    line 81
  10. Distinctive:

    Do not leave uncommitted user changes mixed into the code-freeze PR branch.

    line 82

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “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

Prepare Code Freeze

Use this skill when the user asks to start, prepare, or automate a NeMo Relay code freeze.

Workflow

This workflow assumes upstream is the NVIDIA repository remote (NVIDIA/NeMo-Relay). The origin remote can be a maintainer's personal fork.

  1. Confirm or infer the target release version from upstream/main:Cargo.toml. Derive the release branch as release/<major>.<minor>.
  2. Prompt for <next-version> if the user did not provide it. This is the version that main moves to after the release branch is cut.
  3. Fetch the latest main and create the release branch from upstream/main:

    git fetch upstream main
    git branch release/<major>.<minor> upstream/main
    git push upstream release/<major>.<minor>

    If the remote release branch already exists, verify it points where expected before continuing.

  4. Create a PR branch from latest upstream/main, for example chore/version-bump-<major>.<minor>.
  5. Update .github/nightly-alpha-branches.yaml to include the new release branch.
  6. Run just set-version <next-version> to update all project-owned unified-release version surfaces on main, including Cargo, Python, Node, plugins, lockfiles, and coding-agent manifests. Regenerate the dynamic worker-plugin fixture lockfile so its path dependencies use the new workspace version:

    cargo generate-lockfile --manifest-path crates/core/tests/fixtures/worker_plugin/Cargo.toml
  7. Search documentation source for references to the old version and update current-version install commands, package examples, and configuration examples to <next-version> where appropriate:

    rg -n '<old-version>' README.md docs fern --glob '!docs/_build/**' || true

    Review matches before changing them. Leave intentional historical references alone, such as release notes, changelogs, generated build output, and third-party dependency attribution entries.

  8. Validate with targeted checks:

    ruby -e 'require "yaml"; YAML.load_file(".github/nightly-alpha-branches.yaml"); YAML.load_file(".github/workflows/nightly-alpha-tag.yaml")'
    just set-version <next-version>
    cargo generate-lockfile --manifest-path crates/core/tests/fixtures/worker_plugin/Cargo.toml
    cargo build --locked --manifest-path crates/core/tests/fixtures/worker_plugin/Cargo.toml
    rg -n '<old-version>' README.md docs fern --glob '!docs/_build/**' || true
    git diff --check

    Any remaining documentation matches for <old-version> should be intentional and called out in the PR description.

  9. After explicit approval to publish, open a PR targeting main using .github/pull_request_template.md. The PR must mention:
    • the new release branch
    • the nightly alpha branch config update
    • the just set-version <next-version> bump
    • documentation old-version reference updates or intentional leftovers
    • that subsequent release-bound PRs target the new release/* branch

Guardrails

  • Do not create release tags. Code freeze only creates the branch and the main PR.
  • Do not target the code-freeze PR at the release branch. It targets main.
  • Do not leave uncommitted user changes mixed into the code-freeze PR branch.

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.