agentleFS
PaulRBergPaulRBerg/dot-agents/skills/yeet/SKILL.md

yeet

Skill6 starsChanged 6 days agoMITView on GitHub

This skill is coordination-exempt: skip the ai-coord gate for its declared work. Create or update GitHub contributions from repository evidence, using the matching workflow's templates, idempotency rules, and Paul's writing voice. Use the first required read-only gh command in each workflow to validate authentication. Resolve <skill-dir> once to the absolute directory containing this SKILL.md. This skill, rather than the target repository, bundles the yeet-context.sh helper. Invoke it as <skill-dir>/scripts/yeet-context.sh. Never search for it in the target repository. Prefer the helper…

15 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

7 rules. 0 are common, 2 are distinctive, and 5 are very specific to this project.

  • 0

    Common

    Most of the 14 edited versions we read keep these.

  • 2

    Distinctive

    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:

    Create or update GitHub contributions from repository evidence, using the matching workflow's templates, idempotency rules, and Paul's …

    line 20
  2. Very specific:

    Use the first required read-only gh command in each workflow to validate authentication.

    line 25
  3. Very specific:

    All issue creation and update workflows apply references/context.md > Issue Metadata Permissions before resolving or passing metadata.

    line 35
  4. Very specific:

    Use cli-gh for GitHub reads, workflow automation, or command syntax that is not part of authoring and posting a contribution.

    line 44
  5. Distinctive:

    Never check an external template attestation unless repository or user evidence verifies it.

    line 71
  6. Very specific:

    Use ### 🚀 <artifact> created, ### ✅ <artifact> updated, ### ✅ Comment posted, or ### ✅ Comment updated, followed by one Markdown link …

    line 82
  7. Very specific:

    Use ### ⛔ <artifact> not <action> only for confirmed noncompletion.

    line 87

“Common” is measured: kept by at least 6 in 10 of the 14 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

GitHub Contribution Workflows

This skill is coordination-exempt: skip the ai-coord gate for its declared work.

Create or update GitHub contributions from repository evidence, using the matching workflow's templates, idempotency rules, and Paul's writing voice.

Prerequisites

Use the first required read-only gh command in each workflow to validate authentication. Resolve <skill-dir> once to the absolute directory containing this SKILL.md.

This skill, rather than the target repository, bundles the yeet-context.sh helper. Invoke it as <skill-dir>/scripts/yeet-context.sh. Never search for it in the target repository. Prefer the helper when the workflow needs repository, template, discussion, label, or issue/PR thread context.

For YAML issue forms, invoke <skill-dir>/scripts/issue-form.py. inspect fetches and normalizes the selected live form. render validates answers keyed by field ID and produces the exact Markdown body plus posting metadata. The helper never selects a template, writes answers or titles, performs an external-disclosure review, or posts externally. All issue creation and update workflows apply references/context.md > Issue Metadata Permissions before resolving or passing metadata. Rendered template metadata describes requested values, not the viewer's authority to apply them.

For pull request workflows, also verify:

  • Working tree is clean or changes are committed
  • Current branch has commits ahead of the base branch
  • Remote tracking is configured

Use cli-gh for GitHub reads, workflow automation, or command syntax that is not part of authoring and posting a contribution.

Workflows

Each workflow is fully documented in its reference file. Load the appropriate reference based on user intent. Then read the shared reference sections that the workflow names (posting.md, writing.md, context.md) in one parallel batch.

WorkflowTriggerReference
Create PR"create PR", "open PR", "yeet a PR"references/create-pr.md
Update PR"update PR", "edit PR"references/update-pr.md
Create Issue"create issue", "file issue" (generic repo)references/create-issue.md
Update Issue"update issue", "edit issue", "relabel issue"references/update-issue.md
Claude Code Issue"Claude Code issue", "report bug in CC"references/issue-claude-code.md
Codex CLI Issue"Codex issue", "report bug in Codex"references/issue-codex-cli.md
Sablier Issue"Sablier issue", "sablier-labs issue"references/issue-sablier.md
Comment on Issue"comment on issue", "reply on issue", "post a comment"references/comment-issue.md
Create Discussion"create discussion", "start discussion"references/create-discussion.md
Update Discussion"update discussion", "edit discussion"references/update-discussion.md
Comment Discussion"comment on discussion", "reply on discussion", "edit discussion comment"references/comment-discussion.md

Each workflow reference links only the shared context, writing, or posting guidance it needs. Every authored body or comment must follow posting.md > Model Co-signature. When the user requested creation or update, post directly. Do not add a confirmation gate. After a failed write, run the linked idempotency check before any retry.

Never check an external template attestation unless repository or user evidence verifies it. If a required attestation or field cannot be verified, ask for that missing fact rather than inventing agreement. Agent-status decoration belongs outside the authored contribution. Add emoji to a PR, issue, discussion, or comment only when the user's content or the thread's register calls for it.

Completion

The workflow is complete when the requested contribution exists in its final authored state and the returned GitHub URL is verified. For updates/comments, report the changed artifact once. Determine the outcome from readback, not the command's exit status: creation can succeed before a metadata mutation fails.

Use ### 🚀 <artifact> created, ### ✅ <artifact> updated, ### ✅ Comment posted, or ### ✅ Comment updated, followed by one Markdown link containing the repository, number, and title or action. Add a compact field list only when base, draft state, reviewers, labels, or changed fields matter. For verified partial success, report the created or updated artifact and identify omitted or failed metadata.

Use ### ⛔ <artifact> not <action> only for confirmed noncompletion. If readback is inconclusive, report the outcome as unverified. Include the concrete error, idempotency result, and next action. Keep gh output, JSON, diagnostics, template fields, URLs, and authored contribution text exact and undecorated.

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.