agentleFS
Yingye107Yingye107/practical-work-kit/skills/challenge-my-plan/SKILL.md

challenge-my-plan

Skill1 starsChanged 43 days agoApache-2.0View on GitHub

Stress-test a proposed plan, decision, argument, or approach. Use when the user asks for a second opinion, hidden assumptions, failure modes, tradeoffs, or a better alternative before committing. Assess the plan rather than declaring a delivered artifact complete; check-my-work handles that job.

The file

We split this skill into the rules it gives an agent

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

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Check the best alternative, including a smaller version or keeping the present approach when relevant.

    line 20
  2. Distinctive:

    Use available research tools when an unstable or unfamiliar fact changes the recommendation, and cite what supports the claim.

    line 30

≈ “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

Challenge my plan

Be a useful thinking partner: examine the strongest version of the user's proposal, find what could change the decision, and recommend a practical course. Match their language. Directness is useful; performative negativity is not.

Frame the real decision

Read the proposal and relevant conversation. Identify the desired outcome, constraints, and choice being made. Preserve explicit user priorities. If the plan is sparse, give a provisional assessment with the assumption that matters; ask one focused question only when it changes the recommendation. Do not request a full business case for an everyday decision.

Stress-test selectively

Focus on the few assumptions carrying the outcome. Separate supplied facts, reasonable inference, and things that need checking. Ask what realistic observation would make the plan fail or make another option preferable.

Choose useful lenses for this proposal: feasibility and dependencies, evidence and demand, incentives and human behavior, cost and opportunity cost, reversibility, or failure recovery. Do not mechanically run all lenses. For deeper decisions, use decision-tests.md.

Check the best alternative, including a smaller version or keeping the present approach when relevant. Compare it under the user's actual constraints. Avoid invented probabilities, fake consensus, fixed finding quotas, and generic "needs more research" conclusions.

Distinguish an observed contradiction from a hypothetical failure. Connect each material concern to its trigger, consequence, and the smallest way to check or mitigate it. A plan with no material problem is a valid result; explain why it survives the strongest realistic counterexample you examined.

Make the recommendation actionable

Lead with proceed, change, or stop, with the main reason and appropriate uncertainty. Give the decisive concerns or evidence, then the best workable alternative or adjustment. Usually a few points suffice. End with a concrete next action or a decision test that can change the recommendation.

Respect the requested length. A minor decision may need one paragraph. A consequential choice may need a comparison or a small decision table. Do not force a scorecard, repeated approval rounds, or an exhaustive report.

Use available research tools when an unstable or unfamiliar fact changes the recommendation, and cite what supports the claim. If that evidence is inaccessible, narrow the recommendation and identify the missing fact; do not present intuition as market, legal, medical, or financial evidence.

This skill can work as a single-model structured critique. Only call a review independent when a separate reviewer actually examined the original proposal. Use a second agent only when available, authorized, and worth the cost; give it the proposal and constraints rather than your expected verdict. Tools and delegation are optional, never something to pretend happened.

A critique does not itself authorize editing, spending, publishing, or contacting others. If the user already requested changes, complete the authorized changes and reconsider the affected reasoning.

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.