agentleFS
s977043s977043/river-review/.claude/skills/skill-ops-planner/SKILL.md

skill-ops-planner

Skill7 starsChanged 5 months agoMITView on GitHub

Create an operating policy and roadmap for a live portfolio of Claude Code skills. Use when the user asks to organize, govern, audit, measure, retire, or scale skills already used by a team or repository. Also trigger on "スキルの運用計画", "スキルポートフォリオ", "スキルロードマップ", "スキルを整理して".

The file

We split this skill into the rules it gives an agent

8 rules. About 0 are common, 7 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈7

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If any of the above cannot be determined:

    line 40
  2. Distinctive:

    stop and ask the user for the missing information

    line 42
  3. Distinctive:

    do not proceed to Phase 1 until at least active skills and current storage layout are confirmed

    line 43
  4. Distinctive:

    if no inventory exists at all, output an inventory-first plan and stop

    line 44
  5. Distinctive:

    Create a prioritized roadmap:

    line 129
  6. Very specific:

    Use ${CLAUDE_SKILL_ROOT}/assets/skill-roadmap-template.md as the output structure.

    line 136
  7. Distinctive:

    write abstract policy with no operating rules

    line 186
  8. Distinctive:

    If the team has no current inventory:

    line 194

≈ “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

Pattern declaration

Primary pattern: Pipeline Secondary patterns: Reviewer, Inversion Why: portfolio planning requires strict phase order (inventory → classify → policy → metrics → roadmap) with review checkpoints at each stage. Inversion is needed at Phase 0 to block progress when inventory data is insufficient.

Purpose

Plan how a team should operate a set of skills over time.

Your job is to define:

  • portfolio structure
  • ownership
  • lifecycle stages
  • evaluation policy
  • rollout policy
  • retirement policy
  • review cadence

Phase 0: Portfolio gate

Before planning, identify:

  • active skills
  • target users or teams
  • current pain points
  • risk level by skill category
  • current review process
  • current storage layout
  • current evaluation maturity

If any of the above cannot be determined:

  • stop and ask the user for the missing information
  • do not proceed to Phase 1 until at least active skills and current storage layout are confirmed
  • if no inventory exists at all, output an inventory-first plan and stop

Phase 1: Inventory and pattern segmentation

For each skill, record:

  • owner
  • purpose
  • invocation policy (auto vs manual)
  • risk level
  • primary pattern (Tool Wrapper / Generator / Reviewer / Inversion / Pipeline)
  • secondary patterns
  • current eval maturity
  • review requirement

Output each skill as a structured record:

skill: <name>
owner: <person or team>
purpose: <one line>
invocation: auto | manual
risk: low | medium | high
primary_pattern: <pattern>
secondary_patterns: <patterns or none>
eval_maturity: none | basic | comprehensive
review: <cadence>

Then group the portfolio by:

  • responsibility type (knowledge / workflow / investigation / review / operational)
  • risk type
  • design pattern

Phase 2: Lifecycle policy

Define lifecycle stages:

  1. proposed
  2. experimental
  3. team-approved
  4. production
  5. deprecated
  6. retired

For each stage, define:

  • entry criteria
  • required eval coverage
  • review requirement
  • allowed blast radius
  • logging requirement
  • retirement signals (when to deprecate or retire)

Phase 3: Metrics and governance

Define portfolio metrics:

  • trigger precision
  • task success rate
  • review pass rate
  • rollback rate
  • mean tool count
  • failure recurrence
  • stale skill count

Pattern portfolio metrics:

  • count by primary pattern
  • percentage of Inversion usage where ambiguity is common
  • percentage of Generator skills with template coverage
  • percentage of Reviewer skills with explicit rubric coverage
  • percentage of Pipeline skills with checkpoint validation
  • percentage of Tool Wrapper skills with maintained references

Also define:

  • monthly review cadence
  • change approval policy
  • emergency disable policy
  • documentation freshness policy

Phase 4: Roadmap

Create a prioritized roadmap:

  • fix now
  • standardize next
  • automate later
  • retire soon

Use ${CLAUDE_SKILL_ROOT}/assets/skill-roadmap-template.md as the output structure.

Prioritize by:

  • impact
  • risk
  • maintenance burden
  • adoption level

Phase 5: Review

Check against:

  • ${CLAUDE_SKILL_ROOT}/references/review-default.md
  • ${CLAUDE_SKILL_ROOT}/references/portfolio-policy.md

Verify:

  • policy is concrete enough to operate
  • lifecycle stages have entry rules
  • metrics are measurable
  • ownership is clear
  • retirement criteria exist

Output format

Return exactly these sections:

Current-state summary

Skill inventory model

Lifecycle policy

Governance rules

Metrics dashboard definition

Pattern portfolio

30-day roadmap

90-day roadmap

Risks and open decisions

Anti-patterns

Do not:

  • write abstract policy with no operating rules
  • ignore ownership
  • ignore retirement
  • require eval maturity that the team cannot sustain
  • assume all skills should be auto-invoked

Failure handling

If the team has no current inventory:

  • provide a lightweight bootstrap policy
  • define a minimum viable governance model
  • prioritize inventory, naming, and eval basics first

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.