- State the proposal and the decision it would change.
- Identify at least one plausible alternative or rival explanation, then test which evidence would distinguish it from the preferred view.
- Separate verified evidence, assumptions, unknowns, and claims that need a source. Check that terminology, scope, and success criteria mean the same thing throughout the proposal.
- Review from adversarial, deep, and implementation perspectives: identify neglected stakeholders, boundary conditions, failure modes, dependencies, and whether the proposed mechanism actually produces the claimed outcome.
- Name the highest-impact risk, including privacy, security, cost, quality, or reversal risk when relevant.
- Define what observation would falsify the preferred conclusion and the smallest useful next check.
- Give a calibrated recommendation: proceed, revise, defer, or reject. Do not present a confidence level higher than the evidence supports.
Act Critical Review
Tests a material proposal against alternatives, evidence gaps, risks, and falsifiers. Use when reviewing a design, plan, recommendation, or high-stakes decision.
fabioc-aloha
Updated Act Critical Review by fabioc-aloha · 057f457
npx skillmds@latest add fabioc-aloha/act-critical-review File contents
---name: act-critical-reviewdescription: Tests a material proposal against alternatives, evidence gaps, risks, and falsifiers. Use when reviewing a design, plan, recommendation, or high-stakes decision.---1. State the proposal and the decision it would change.2. Identify at least one plausible alternative or rival explanation, then test which evidence would distinguish it from the preferred view.3. Separate verified evidence, assumptions, unknowns, and claims that need a source. Check that terminology, scope, and success criteria mean the same thing throughout the proposal.4. Review from adversarial, deep, and implementation perspectives: identify neglected stakeholders, boundary conditions, failure modes, dependencies, and whether the proposed mechanism actually produces the claimed outcome.5. Name the highest-impact risk, including privacy, security, cost, quality, or reversal risk when relevant.6. Define what observation would falsify the preferred conclusion and the smallest useful next check.7. Give a calibrated recommendation: proceed, revise, defer, or reject. Do not present a confidence level higher than the evidence supports.
fabioc-aloha/act_skills_for_scout/tree/main/packages/act-skills-for-scout-v1.0.0/act-critical-review commit 057f457bfd
Frequently asked questions
Run npx skillmds@latest add fabioc-aloha/act-critical-review in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.
Tests a material proposal against alternatives, evidence gaps, risks, and falsifiers. Use when reviewing a design, plan, recommendation, or high-stakes decision. It is listed under Coding & Dev Tools on SkillMD.
This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.
This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.
Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.
fabioc-aloha (@fabioc-aloha) published this skill. Their other Agent Skills are listed on their SkillMD profile.