---
name: fu-action-ranker
description: Rank Future Union action ideas against movement values, impact, sourceability, safety, and review effort before they become public board cards.
---

# FU Action Ranker

Use this skill when a Future Union mission asks you to shape or score a proposed action.

## Workflow

1. Restate the proposed action in one sentence.
2. If `INTAKE_IDEA.md` or `context/intake-idea.json` exists, treat it as unverified raw Intake context.
3. Identify the concrete output it would produce.
4. Check whether the idea has public evidence or a clear way to gather it.
5. Score the proposal from 1-5 on each criterion:
   - future upside
   - public usefulness
   - sourceability
   - actionability
   - leverage per hour
   - movement fit
   - safety / reputation risk
   - review effort
6. Recommend exactly one queue outcome:
   - promote to Open now
   - needs shaping
   - needs backing
   - hold
   - reject

## Ranking Standard

- Prefer actions that create evidence, pressure, distribution, useful tools, or clear local proof.
- Prefer narrow outputs a real person or agent can complete in one sitting.
- Prefer source-backed claims over slogans, vibes, or factional point-scoring.
- Mark safety, privacy, legal, reputational, and review-load risks plainly.
- Keep contribution credit separate from impact scoring. A useful contributor does not make an unsafe idea safe.
- If the proposal is promising but unclear, recommend `needs shaping`, not `promote`.
- If the proposal requires money, access, specialist review, or repeated compute before it can run, recommend `needs backing`.

## Output

Return Markdown using `ACTION_PROPOSAL.md`.

Include:

- proposed action
- problem
- target user or audience
- source trail
- proposed output
- time box
- expected impact
- ranking table
- risks and caveats
- first reviewer question
- recommended queue
- contribution credit note, if `CREDIT.md` is present

## Never Do

- Do not publish, move, or relabel board cards yourself.
- Do not contact third parties.
- Do not expose raw Intake contact details or private notes in the public proposal.
- Do not invent evidence, scores, or source links.
- Do not recommend public action where the safety or evidence risk is unresolved.
- Do not turn every idea into an action. Reject weak or unsafe proposals cleanly.