agentleFS
yicheng47yicheng47/runner/.agents/skills/feature/SKILL.md

feature

Skill490 starsChanged 7 days agoMITView on GitHub

File, list, or manage feature issues, and write a feature's spec when its development starts

71 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use bugfix, adopt and 9 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.
  • team-attentionSkill

    bugfix

    Root cause based one-shot bug fix. debugger diagnosis → requirements.md generation → /execute. /bugfix "error description" Full investigation pipeline: debugger + gap-analyzer + standard verify. QA suggestion after successful fix.

    Also in 7 projects that use this

  • codebycoreySkill

    adopt

    Adopt an existing plan from conversation context into delta-spec format. Use when importing a plan, adopting planning that already happened, capturing plan mode output, or bringing in prior design work.

    Also in 4 projects that use this

  • Google CloudSkill

    fix

    Plan a remediation strategy after a test failure.

    Also in 10 projects that use this

  • marylinSkill

    recover

    Recover work from a crashed or interrupted session. Use when Claude crashed mid-task, after an unexpected exit, when you see 'stale sessions detected', or when you lost context and need to figure out what was happening.

    Also in 3 projects that use this

  • aegiswizardSkill

    context

    Also in 4 projects that use this

  • AWSSkill

    map

    Map verified service capabilities against customer-provided control framework and threat model. Produces mapping-results.json, controls-catalog.md, and framework-mapping.md. Supports pluggable frameworks: CCMv4, NIST 800-53, CIS Benchmarks, ISO 27001, etc. Split into 4 focused sub-skills for better accuracy; this skill orchestrates them.

    Also in 4 projects that use this

The file

Show its rules and text

We split this skill into the rules it gives an agent

14 rules. 0 are common, 6 are distinctive, and 8 are very specific to this project.

  • 0

    Common

    Most of the 65 edited versions we read keep these.

  • 6

    Distinctive

    This author's own choices. Worth a look.

  • 8

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Ask the user to describe the feature (motivation, scope, key decisions) unless they already have.

    line 19
  2. Very specific:

    Create the issue with labels feature, the chosen P0/P1/P2/P3, the readiness label, if any, and backlog when it has no milestone:

    line 23
  3. Very specific:

    If the user asks for closed/shipped features too, repeat with --state all and add a State column.

    line 37
  4. Distinctive:

    Write the spec when development of a feature starts.

    line 40
  5. Very specific:

    Create docs/features/{issue}-{slug}.md, pre-populated from the issue body and its discussion: a header linking the tracking issue, then …

    line 43
  6. Very specific:

    Add the spec's line under Active in docs/features/README.md.

    line 44
  7. Very specific:

    Leave its spec where it is.

    line 51
  8. Very specific:

    Remove any existing P-label on the issue, then add the new one: gh issue edit <number> --remove-label P0 --remove-label P1 --remove-label …

    line 56
  9. Very specific:

    Remove both readiness labels, then add the new one unless it is none: gh issue edit <number> --remove-label agent-ready --remove-label …

    line 64
  10. Distinctive:

    Put an existing feature on a milestone or in the backlog.

    line 69
  11. Very specific:

    Update the label as the issue moves: remove needs-design once the design is signed off or the questions are settled, and add agent-ready if …

    line 94
  12. Distinctive:

    Every feature gets exactly one priority label.

    line 98
  13. Distinctive:

    When in doubt between two levels, pick the lower-urgency one and say why.

    line 105
  14. Distinctive:

    Do not commit or push unless the user explicitly asks.

    line 115

“Common” is measured: kept by at least 6 in 10 of the 65 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

Feature Management

A feature is a GitHub issue with the feature label until someone starts building it. Filing, prioritizing and scheduling a feature touch only GitHub: nothing is written to the repo and nothing is committed. Its spec in docs/features/ is written when development starts, on the feature's branch, and reaches main in the same pull request as the code, together with any impl plan or mission brief.

Usage

/feature <action> [args]

Actions

new <name>

File a new feature as an issue.

  1. Ask the user to describe the feature (motivation, scope, key decisions) unless they already have.
  2. Pick a priority (see Priority below). If the user didn't state one, propose one and confirm before filing. Don't file unlabeled.
  3. Pick a readiness label (see Readiness below) and propose it together with the priority. A feature with new UI or open questions is needs-design.
  4. Schedule it (see Scheduling below): set a milestone only when the user names one; otherwise add backlog.
  5. Create the issue with labels feature, the chosen P0/P1/P2/P3, the readiness label, if any, and backlog when it has no milestone:
    • Title: feat: <short description>
    • Body: Motivation, Scope, and Open questions when there are any. The issue holds the design discussion until development starts.
    • Command shape: gh issue create --label feature --label P1 [--label needs-design] --title "…" --body "…" (--milestone 0.16 | --label backlog)
  6. Report the issue URL, priority, readiness and milestone or backlog. Do not create a spec or edit any doc.

list

Show all features, sorted by priority.

  1. Fetch open features with priority + metadata as JSON so they can be sorted: gh issue list --label feature --state open --limit 50 --json number,title,labels,milestone,createdAt,assignees
  2. List the spec files in docs/features/ (excluding README); most features have none until development starts.
  3. Sort the rows by priority (P0 first, then P1, P2, P3, then unlabeled-by-priority last). Within a priority bucket, sort by issue number ascending.
  4. Present a combined view: Priority, Readiness (agent-ready, needs-design or —), feature name, # (issue), Milestone (backlog for a backlog issue), spec file (if one exists), Created. Issues without a P-label show — in priority, and issues with neither a milestone nor backlog show — in Milestone; add a callout asking the user to triage either.
  5. If the user asks for closed/shipped features too, repeat with --state all and add a State column.

spec <issue-number>

Write the spec when development of a feature starts.

  1. Work in the feature's branch worktree (AGENTS.md, Worktrees), never in the root checkout on main. If the branch does not exist yet, create it as AGENTS.md describes once the user has asked to start the work.
  2. Create docs/features/{issue}-{slug}.md, pre-populated from the issue body and its discussion: a header linking the tracking issue, then Motivation, Scope, Implementation Phases, Verification. Leave priority and milestone out; they live on the issue.
  3. Add the spec's line under Active in docs/features/README.md.
  4. The spec is committed on the branch and lands with the code. A crew mission folds it into its commits like the brief.

close <issue-number>

Close a feature that shipped or was dropped.

  1. Close the GitHub issue: gh issue close <number>, adding --reason "not planned" when it was dropped.
  2. Leave its spec where it is. The release's docs sweep archives specs (release skill, sweep).

prioritize <issue-number> <P0|P1|P2|P3>

Set or change the priority of an existing feature.

  1. Remove any existing P-label on the issue, then add the new one: gh issue edit <number> --remove-label P0 --remove-label P1 --remove-label P2 --remove-label P3 --add-label <priority> (Removing all four is safe — gh ignores remove-label for labels not present.)
  2. Confirm the new priority.

ready <issue-number> <agent-ready|needs-design|none>

Set or change an existing feature's readiness.

  1. Remove both readiness labels, then add the new one unless it is none: gh issue edit <number> --remove-label agent-ready --remove-label needs-design [--add-label <readiness>]
  2. Confirm the new readiness.

schedule <issue-number> <milestone|backlog>

Put an existing feature on a milestone or in the backlog.

  1. For a milestone: gh issue edit <number> --milestone <milestone> --remove-label backlog. For the backlog: gh issue edit <number> --remove-milestone --add-label backlog.
  2. Confirm the new milestone or backlog. docs/roadmap.md mirrors the split at its next refresh.

Labels

  • feature — all feature issues use this label.
  • bug — for bug reports (not managed by this skill).
  • P0 / P1 / P2 / P3 — priority, exactly one per issue.
  • agent-ready / needs-design — readiness, at most one per issue (see Readiness below).
  • backlog — no release commitment; see Scheduling below.

Scheduling

Every open issue has exactly one of a milestone or the backlog label. A milestone is a release commitment; backlog means the issue is triaged but has none. Setting a milestone removes backlog, and taking an issue off its milestone without closing it adds backlog. Priority and scheduling are separate: a P1 can sit in the backlog when its development is deliberately later.

Readiness

A readiness label says what has to happen before development can start. Many features get neither.

  • needs-design — development cannot start yet: the feature adds or changes UI that has no signed-off Pencil frame, it has open questions the user must decide, or it spans several surfaces and needs a spec or plan first. Most new features start here.
  • agent-ready — small and fully specified: nothing is left to decide, no UI needs drawing, and an agent can build and verify it in one pass. Build it inline, not as a crew mission.
  • Neither — specified but substantial work for a crew mission, or an evaluation or investigation rather than a build.

Update the label as the issue moves: remove needs-design once the design is signed off or the questions are settled, and add agent-ready if what is left is small.

Priority

Every feature gets exactly one priority label. Rubric:

  • P0 — Required for the next ship; nothing else moves until this lands. Rare for features.
  • P1 — Wanted this cycle; blocks an active user workflow or has a stakeholder commitment.
  • P2 — Real product win, but no urgency. Pick up when the P1 queue is clear.
  • P3 — Idea / nice-to-have / "if we ever revisit X." OK to sit indefinitely; closing as wontfix later is fine.

When in doubt between two levels, pick the lower-urgency one and say why; over-labeling P0/P1 dilutes the signal.

Conventions

  • A spec is named after its tracking issue, {issue}-{slug}.md, with a lowercase kebab-case slug.
  • Specs, impl plans (docs/impls/) and mission briefs (docs/impls/briefs/) are written on the feature's branch and land with its code. Design files are the exception: .pen files are settled on main first (AGENTS.md, Engineering Conventions).
  • At the release's docs sweep, retain specs in docs/features/archive/ only when their decisions, compatibility constraints or implementation context still help. Obsolete or duplicate proposals can be pruned after preserving useful unique content and fixing references; issues and git history are enough for many dropped ideas. The implementation is the source of truth.

Notes

  • Do not commit or push unless the user explicitly asks.
  • A spec always links its GitHub issue URL.

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.