Critical Ideation (Adversarial)

Purpose

Challenge the idea to surface hidden risks, weak assumptions, and better alternatives before technical design. Run as a sub-agent and return a draft section to the orchestrator; do not write the plan artifact directly.

When to use

• •After problem definition is complete
• •Before technical planning

Required outputs (Challenge Packet)

• •2 concrete weaknesses (specific scenario -> impact)
• •1 end-to-end failure mode
• •Assumptions -> tests (pass/fail)
• •Ranked risks with owners + status
• •Alternatives considered (include one disliked alternative)
• •At least 1 measurable milestone
• •Decision boundary if a real fork exists
• •Draft section content for ## Challenge Artifacts

Sub-agent output contract

Return a single block in this shape:

DraftSection:
<exact section content for ## Challenge Artifacts (must include the section header)>

Checklist:
- <criterion>: Pass | Fail

Questions:
- <if blocked>

Notes:
- <optional risks/assumptions/tests updates>

Malformed output handling

• •If you cannot produce the exact section header or required fields, return Questions explaining what is missing and leave DraftSection as N/A.

Success criteria (gate: FeatureClarity)

• •Weaknesses are concrete and distinct
• •At least 1 failure mode is end-to-end
• •Top assumptions have tests
• •Top risks have owners + status
• •Alternatives include one disliked option
• •Evaluation criteria is present in the plan

Process

1. •Restate the problem definition neutrally.
2. •Identify weaknesses and failure mode.
3. •Convert assumptions to tests.
4. •Rank risks and assign owners/status.
5. •List alternatives (include disliked alternative).
6. •Define milestone + evidence.
7. •Add decision boundary if needed.

Output template

Use this exact template:

### Challenge Packet
#### 1) What I think you’re trying to do (1–3 bullets)
- ...

#### 2) Two concrete weaknesses (must be concrete)
- W1:
- W2:

#### 3) One failure mode (end-to-end)
- FM1: <failure> — detection — mitigation/prevention

#### 4) Assumptions → tests (pass/fail)
- A1: <assumption>
  - Test:
  - Pass/Fail criteria:
  - Status: Untested | Tested | Accepted | Deferred
- A2: ...

#### 5) Ranked risks (with owners + status)
- R1 (High): <risk>
  - Mitigation:
  - Owner:
  - Status: Mitigated | Tested | Accepted | Deferred
  - Trigger (if deferred):

#### 6) Alternatives considered (including one you dislike)
- Alt A:
  - Why it might win:
  - Why you’d reject it:

#### 7) Milestone(s) + evidence
- Milestone:
  - Evidence:

#### 8) Decision boundary (ONLY if required)
- Decision needed:
  - A) ...
  - B) ...
  - C) ...
Recommended default: <A/B/C> (why)

UX rules

• •Be direct; avoid praise or marketing language.
• •Do not invent context; challenge only what is provided.
• •If the user rubber-stamps, require specifics or run the human Q/A loop.
• •Deferred requires DR + trigger (hard rule): do not mark any assumption/test/risk “Deferred” unless you also provide a DR-xxx and an explicit trigger for revisit.