agentleFS
grafanagrafana/gcx/.agents/skills/propose-feature/SKILL.md

propose-feature

Skill819 starsChanged 18 days agoApache-2.0View on GitHub

Turn an idea or rough notes into a reviewable gcx feature proposal through a short interview. Use when asked to propose, draft, or file a feature or improvement.

1 other project uses a version of this.

Adopted by

Most recently taken up by

  • Other versions1 project uses a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 1

    1. mediajunkie/piper-morgan-product★ 8same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository8 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

10 rules. About 0 are common, 8 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈8

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

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

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

    Read the contribution workflow for when a proposal is needed and what follows acceptance.

    line 10
  2. Distinctive:

    Check VISION.md for product fit and CONSTITUTION.md for invariants.

    line 18
  3. Distinctive:

    Read relevant code and existing command help to understand the current behavior and nearby capabilities.

    line 19
  4. Distinctive:

    Search existing issues and PRs for duplicates or decisions to reuse.

    line 20
  5. Distinctive:

    Stop interviewing when the problem, proposed UX and checkable acceptance criteria are clear.

    line 29
  6. Distinctive:

    Write a title that names the outcome.

    line 52
  7. Distinctive:

    Check every existing command, flag and file path against the code or command help.

    line 54
  8. Distinctive:

    When filing is authorized:

    line 62
  9. Very specific:

    Create the issue with gh issue create --body-file, applying the form's label.

    line 64
  10. Distinctive:

    Read back the issue body and metadata, then report its URL and any other authorized changes.

    line 66

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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

Propose Feature

Produce one proposal that a reader with no access to the conversation can understand and assess. Use the common Feature proposal form for requests, improvements and proposed commands. This skill ends with a reviewable draft or an authorized, filed issue.

Read the contribution workflow for when a proposal is needed and what follows acceptance. Use contribute when selecting the contribution route or preparing implementation. Follow that policy; a user may also request a proposal draft for work whose route does not require one.

The repository is public. Follow the anonymization rules in the GitHub Issues section of CLAUDE.md. Use public evidence or independently developed material suitable for public review. Keep restricted research outside the checkout; removing links or attribution does not make internal findings publishable.

Ground the idea

Establish the concrete user problem and investigate enough context to make the direction credible:

  • Check VISION.md for product fit and CONSTITUTION.md for invariants.
  • Read relevant code and existing command help to understand the current behavior and nearby capabilities.
  • Search existing issues and PRs for duplicates or decisions to reuse.
  • Investigate relevant public APIs or dependencies where they affect feasibility. Scale research to the proposal; state unknown API readiness, permissions, data availability or other feasibility gaps explicitly.

Separate observed facts, user-reported needs and proposed behavior. When evidence contradicts the idea or an invariant prevents it, explain the conflict and reshape the proposal before drafting. A proposal can carry unresolved feasibility questions for maintainer review.

Interview briefly

Reuse decisions and evidence already available. Ask only what changes the proposal: who has the problem, the scenario that shows it, where scope stops and what completion looks like. Use a small batch of questions with recommended answers and tradeoffs where useful; continue independent research while answers are pending.

Stop interviewing when the problem, proposed UX and checkable acceptance criteria are clear. Record unresolved design choices as open questions.

Draft

Cover the form's three required sections. Add optional sections only where they help review:

## Problem
<who encounters what today, and why current behavior or workarounds are insufficient>

## Proposed UX
<the user experience, illustrated at a scale appropriate to the change>

## Acceptance criteria
- [ ] <observable outcome a reviewer can check>

## Scope and open questions
<boundaries, alternatives and unresolved feasibility or design questions>

## Implementation notes
<relevant constraints, verified dependencies and willingness to implement, where useful>

Write a title that names the outcome. Use a short before/after for a small improvement; add command, output or error examples when they make a larger proposal concrete. Issue #1469 illustrates a detailed UX proposal, rather than a required length or design checklist.

Check every existing command, flag and file path against the code or command help. Mark new commands and behavior as proposed. Each acceptance criterion must be checkable without the conversation. Label feasibility assumptions and user-reported claims so readers can distinguish them from verified facts.

Review and file

Show the full draft with its title, issue type (Feature), labels (action/needs-triage) and any proposed assignees. Revise until it is ready for approval.

Publish only when the user's instructions authorize the concrete issue and destination; reuse authorization already given for the same action. Until then, hand back the private draft. Apply the same authorization boundary to assignments, links and changes to other shared records.

When filing is authorized:

  • Create the issue with gh issue create --body-file, applying the form's label.
  • Set the type to Feature with the GraphQL updateIssueIssueType mutation, looking up the existing type ID from the organization's issueTypes. Preserve organization-wide type administration.
  • Read back the issue body and metadata, then report its URL and any other authorized changes.

The next step is maintainer review. After acceptance, follow the contribution guide for the new-feature RFC requirement and product-team exception. Reuse accepted issue decisions in the RFC; developers choose their own implementation tools and planning approach.

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.