mcpskills.net — Curated AI Skills, MCPs & Agents

mcpskills.net

A confident answer is not a correct one. Doubt-driven development materializes a fresh-context reviewer biased to disprove before any non-trivial output stands.

When to Use

A decision is non-trivial when at least one of these is true:

It introduces or modifies branching logic
It crosses a module or service boundary
It asserts a property the type system cannot verify
Its correctness depends on context the future reader cannot see
Its blast radius is irreversible (production deploy, data migration, public API)

The Process

Doubt cycle:
- [ ] Step 1: CLAIM — wrote the claim + why-it-matters
- [ ] Step 2: EXTRACT — isolated artifact + contract, stripped reasoning
- [ ] Step 3: DOUBT — invoked fresh-context reviewer with adversarial prompt
- [ ] Step 4: RECONCILE — classified every finding against the artifact text
- [ ] Step 5: STOP — met stop condition

Step 1: CLAIM

Name the decision in two or three lines:

CLAIM: "The new caching layer is thread-safe under the
        read-heavy workload described in the spec."
WHY THIS MATTERS: a race here corrupts user data and is
                  hard to detect in QA.

Step 2: EXTRACT

A fresh-context reviewer needs the artifact and the contract, not the journey.

Code: the diff or the function — not the whole file
Decision: the proposal in 3–5 sentences plus constraints
Assertion: the claim plus the evidence

Strip your reasoning. If you hand over conclusions, you'll get back validation of your conclusions.

Step 3: DOUBT

The reviewer's prompt must be adversarial:

Adversarial review. Find what is wrong with this artifact.
Assume the author is overconfident. Look for:
- Unstated assumptions
- Edge cases not handled
- Hidden coupling or shared state
- Ways the contract could be violated
- Existing conventions this might break
- Failure modes under unexpected input

Do NOT validate. Do NOT summarize. Find issues, or state
explicitly that you cannot find any after thorough examination.

ARTIFACT: <paste artifact>
CONTRACT: <paste contract>

Pass ARTIFACT + CONTRACT only. Do NOT pass the CLAIM. Handing the reviewer your conclusion biases it toward agreement.

Step 4: RECONCILE

For each finding, classify in this precedence order:

Contract misread — reviewer flagged because CONTRACT was unclear. Fix the contract first.
Valid + actionable — real issue requiring a change. Change it, re-loop.
Valid trade-off — issue is real but cost of fixing exceeds cost of accepting. Document explicitly.
Noise — reviewer flagged something actually correct under context it didn't have.

Step 5: STOP

Stop when:

Next iteration returns only trivial or already-considered findings, or
3 cycles completed (escalate to user), or
User explicitly says "ship it"

Verification

After applying doubt-driven development:

[ ] Every non-trivial decision was named explicitly as a CLAIM
[ ] At least one fresh-context review per non-trivial artifact
[ ] The reviewer received ARTIFACT + CONTRACT — NOT the CLAIM
[ ] The reviewer's prompt was adversarial ("find issues")
[ ] Findings were classified against the artifact text
[ ] A stop condition was met (trivial findings, 3 cycles, or user override)
[ ] In interactive mode, cross-model was explicitly offered