agentleFS
redhat-developerredhat-developer/rhdh-skills/skills/plugins/rhdh-pr-create/SKILL.md

rhdh-pr-create

Skill20 starsChanged 20 days agoApache-2.0View on GitHub

Publishes verified changes from rhdh-plugins or community-plugins: detect the repository and affected workspaces, run the repository build pipeline, create package changesets, stage generated files safely, create a signed-off commit and branch, push, open a GitHub pull request, upload optional bug-fix recordings, and link Jira or GitHub issues. Use for raise PR, create or open a plugin PR, push verified plugin changes, or publish a verified change another skill handed off.

  • Commits and pushes

The file

We split this skill into the rules it gives an agent

7 rules. About 0 are common, 3 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Run gh auth status and inspect branch, status, staged diff, and upstream default branch.

    line 24
  2. Distinctive:

    If the calling skill supplied a change handoff, validate its repository and file list, issue reference, recordings, and verification …

    line 26
  3. Very specific:

    Follow workflows/create-pull-request.md sequentially.

    line 29
  4. Very specific:

    Only published plugin source paths need changesets.

    line 48
  5. Distinctive:

    Never fabricate issue data, recording URLs, CI results, or reviewer evidence.

    line 50
  6. Distinctive:

    When no handoff is supplied, derive these fields from the staged diff and ask only for unresolved issue context or release intent.

    line 66
  7. Very specific:

    Read-only inspection, builds, and draft construction do not approve a write.

    line 71

≈ “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

RHDH Pull Request

Own publication after implementation is verified. Accept either the current staged checkout or a change handoff from the calling skill; do not diagnose or modify product code.

Start here

  1. Load references/repo-profiles.md and identify the canonical upstream from all remotes.
  2. Run gh auth status and inspect branch, status, staged diff, and upstream default branch.
  3. If the calling skill supplied a change handoff, validate its repository and file list, issue reference, recordings, and verification evidence against the checkout. Treat the checkout as authoritative and report mismatches.
  4. Follow workflows/create-pull-request.md sequentially. There is no auto-approve mode: every external write is stated in full and approved as that stated set before it runs.

Boundaries

  • This skill stages, formats, validates, creates changesets, commits, pushes, opens the PR, uploads supplied recordings, and updates linked GitHub issues. Jira reads belong to /rhdh-jira-api and Jira writes to /rhdh-jira-update. Invoke either by name and consume what it reports; never call Jira directly from here.
  • GitHub issue reads belong to /rhdh-forge, which returns the same issue detail with source: github. Load references/github-input.md when the request supplies an issue URL or number.
  • It does not implement fixes or features. If validation exposes a product-code failure, stop and hand the failure back to the skill that produced the change, usually /rhdh-plugin-bug-fix, with the failing command and its output.
  • Pre-existing dirty or untracked files are outside the publication set unless the user explicitly identifies them as part of the change.
  • Only published plugin source paths need changesets; private packages/*, dev apps, tests, fixtures, and stories do not.
  • Never fabricate issue data, recording URLs, CI results, or reviewer evidence.

Change handoff

A skill that produced a verified change may hand it off instead of leaving it staged; /rhdh-plugin-bug-fix is the usual producer. The handoff arrives in the conversation and names the change summary, the file list making up the change, the issue reference, the optional before and after recordings, and the verification that was run with its result.

The file list is the change set. The producing skill does not stage, so those paths arrive unstaged or untracked: verify them against the working tree, keep them out of the pre-existing baseline, and stage exactly them alongside the build-generated files at the workflow's staging gate. An empty index is a stop condition only when no handoff was supplied.

When no handoff is supplied, derive these fields from the staged diff and ask only for unresolved issue context or release intent.

External writes

Read-only inspection, builds, and draft construction do not approve a write. A push, recording upload, PR creation, or GitHub issue update is an external write: invoke the named skill mutation-gate and follow it rather than restating the gate here.

State every write this skill owns with its exact target, exact command, and a preview of the change — the git push, the GitHub Contents upload, the pull-request creation, the GitHub issue comment — get approval for that stated set, execute only that set, and then report the outcome of every operation, including the ones that were skipped. Stage exactly the approved paths, taken from the handoff's file list when a handoff supplied them. If an earlier operation produces material a later one needs, such as an uploaded recording URL used in the PR body, report the outcomes of the first set, then state and get approval for the next one. Reject legacy --a as unsupported. Each reported outcome also names the changed resource or URL, its verification, and any remaining recovery action.

Completion

Report the PR URL, the repository, the head branch and commit SHA, the generated changesets, the uploaded recording URLs when recordings were supplied, every issue update, and the outcome of every external write. Take exact URLs and SHAs from command output rather than reconstructing them. A created PR is not complete until its URL is captured and the requested upload gates have either succeeded or been reported for manual action. If an external update fails, keep the successful PR result and report the failed update rather than hiding it.

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.