agentleFS
m2ai-mcp-serversm2ai-mcp-servers/claude-skills/skills/counterargument-stress-test/SKILL.md

counterargument-stress-test

Skill0 starsChanged 7 months agoMITView on GitHub

Takes any strategic argument, business case, or thesis and systematically generates and addresses the N strongest counterarguments, forcing rigorous thinking before publishing or deciding.

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 3 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    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:

    Use when the user says "stress test this argument", "devil's advocate", "what are the counterarguments", "poke holes in this", "is this …

    line 12
  2. Distinctive:

    If the user doesn't specify N, default to 5 counterarguments.

    line 23
  3. Distinctive:

    If WEAK REBUTTAL, flag this as a genuine vulnerability in the original argument.

    line 45

≈ “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

Counterargument Stress-Tester

Takes a position, thesis, business case, or strategic argument and pressure-tests it by generating the strongest possible counterarguments, then evaluating whether the original position survives.

Trigger

Use when the user says "stress test this argument", "devil's advocate", "what are the counterarguments", "poke holes in this", "is this argument solid", "steelman the other side", or provides a thesis/position and asks for critical review.

Phase 1: Intake

Accept the argument. This can be:

  • A written thesis or position statement
  • A business case or proposal
  • A blog post draft or newsletter
  • A strategic decision ("we should switch to X")
  • A spec-gap-detector output that needs the strategic layer tested

If the user doesn't specify N, default to 5 counterarguments. Accept any N from 3-10.

Restate the argument back in one sentence to confirm understanding before proceeding.

Phase 2: Generate Counterarguments

For each counterargument:

  1. State it as a steel-man. The strongest, most charitable version of the opposing view. No straw-men.
  2. Identify the type:
    • EMPIRICAL -- "The data doesn't support this" (cites evidence)
    • LOGICAL -- "The reasoning has a flaw" (identifies fallacy or gap)
    • CONTEXTUAL -- "This doesn't apply here" (scope/timing/audience mismatch)
    • PRACTICAL -- "This won't work in practice" (execution, cost, adoption barriers)
    • MORAL/ETHICAL -- "This shouldn't be done" (values, externalities, fairness)
  3. Rate the strength: WEAK (easily dismissed) / MODERATE (requires a response) / STRONG (potentially fatal to the argument)
  4. Provide evidence or reasoning for the counterargument -- don't just assert it, support it.

Phase 3: Rebuttal Pass

For each counterargument, draft a rebuttal from the perspective of the original argument:

  • Does the rebuttal hold? Rate: STRONG REBUTTAL / PARTIAL REBUTTAL / WEAK REBUTTAL
  • If WEAK REBUTTAL, flag this as a genuine vulnerability in the original argument.

Phase 4: Verdict

Produce a summary assessment:

## Stress Test Results

**Argument:** [one-line restatement]
**Counterarguments tested:** N
**Breakdown:** X WEAK | Y MODERATE | Z STRONG

### Vulnerabilities Found
[List any counterarguments where the rebuttal was PARTIAL or WEAK -- these are real gaps]

### Verdict
[One of:]
- ROBUST -- argument survives all major challenges
- CONDITIONALLY SOUND -- holds under stated conditions, fragile outside them
- NEEDS REVISION -- 1-2 significant gaps that should be addressed before publishing/deciding
- FATALLY FLAWED -- a STRONG counterargument with a WEAK rebuttal undermines the core thesis

### Suggested Revisions
[If NEEDS REVISION or FATALLY FLAWED: specific changes to strengthen the argument]

Phase 5: Output

Present the full analysis, then offer:

  • "Want me to rewrite the argument incorporating these findings?"
  • "Want to go deeper on any specific counterargument?"
  • "Want me to run spec-gap-detector on the revised version?"

Verification

A good stress test:

  • Has at least one MODERATE or STRONG counterargument (if all are WEAK, the test wasn't rigorous enough -- try harder)
  • Steel-mans genuinely -- the counterarguments should make the user uncomfortable, not just check a box
  • Rebuttals are honest -- marking a weak rebuttal as strong defeats the purpose
  • The verdict matches the evidence (don't soften FATALLY FLAWED to spare feelings)

Source

Extracted from Nate Kadlac newsletter (2026-03-29) -- "Executive Briefing: 33% of the world's helium supply just went offline" -- explicit counterargument addressing methodology as a standalone skill.

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.