deanpeters/product-manager-antagonist-skills / skills/pant-artifact-pressure-test
pant-artifact-pressure-test
VerifiedPressure-test a product artifact, argument, or proposal when no specialist PANT scenario fits. Use for general grounding, contradictions, evidence, and readiness; not work clearly served by a specialist.
PANT Artifact Pressure Test
Purpose
Provide structured dissent for a product artifact or argument when no specialist scenario materially improves the review. Ground critique in supplied material, prioritize consequential weaknesses, convert findings to action, and route only when specialization is useful.
When to Use
Use for product proposals, decision memos, presentations, demonstrations, research summaries, plans, narratives, and other artifacts whose specific scenario is unclear or outside the current catalog.
When Not to Use
Prefer the roadmap, strategy, business-case, market-sizing, tactical-plan, or executive-decision specialist when its decision is central. Use pant-meeting-rehearsal when the primary job is conversational practice rather than artifact review. Do not pretend this generalist provides specialist professional approval.
Supported Modes and Entry Paths
- Feedback: Default. Apply the bundled feedback contract.
- Dialogue: Follow up on one unresolved claim or rehearse defense of the artifact one question at a time.
- Direct: Start from the supplied material immediately.
- Guided: Explain the claim-evidence-decision relationship and ask only what changes the review.
- Best effort: State audience, decision, and consequence assumptions; review with reduced confidence.
Decision Being Prepared For
Prepare the artifact to inform, recommend, request, align, or support a consequential product decision. If no decision exists, clarify the intended audience response without manufacturing an approval request.
Typical Audiences
Product teams, cross-functional partners, leaders, executives, customers, sponsors, portfolio groups, and other decision participants.
Inputs
Best input is the artifact plus audience, intended response, stakes, and known evidence. Minimum input is the artifact or a description of its argument. Everything supplied counts as context.
Direct example:
Use
pant-artifact-pressure-teston this proposal. Identify what survives, the three most consequential weaknesses, unsupported claims, contradictions, and the smallest revisions that improve readiness.
Sparse example:
I have a product document that feels weak, but I do not know which review I need.
Quick Start
- Identify artifact type, intended audience response, and consequence.
- Read general artifact logic.
- Extract load-bearing claims and classify their support.
- Test contradictions, missing context, alternatives, and actionability.
- Route only if a specialist would materially improve the work.
Review Workflow
- Name the job. State what the artifact is trying to cause the audience to understand, decide, or do.
- Find what survives. Preserve supported claims, useful structure, honest limitations, and clear decisions.
- Map the argument. Identify claims, evidence, assumptions, forecasts, commitments, and unknowns.
- Test consequence. Prioritize weaknesses by likely effect on the intended decision or action.
- Test contradictions. Inspect internal inconsistency, evidence that undermines claims, and mismatches among audience, ask, and detail.
- Test omissions. Examine relevant alternatives, tradeoffs, ownership, adoption, risk, and uncertainty without forcing every lens.
- Choose depth. Continue generally when sufficient; route or combine when a specialist decision dominates.
- Convert to action. Recommend revisions, evidence, discovery, mitigations, or decision conditions.
- Avoid rewriting as concealment. Make the critique visible before offering replacement language.
- Issue a contextual verdict. Judge readiness for the identified audience response.
Dialogue Protocol
Ask one question about the highest-priority unresolved claim or contradiction. Follow the user's actual answer and update the concern. Honor coaching, pressure, and debrief controls. Do not cosplay a stakeholder when role context is absent; challenge the argument directly.
Output
Use the shared feedback order and add an Artifact decision map:
- Artifact job, audience, and intended response
- Load-bearing claims and evidence classifications
- What survives scrutiny
- Blocking and significant concerns
- Contradictions, tradeoffs, and missing context
- Specialist routing recommendation, if material
- Prioritized revisions and evidence actions
- Contextual readiness verdict
Use template.md for a reusable review.
Failure Modes
- Producing generic comments applicable to any document
- Reviewing prose while ignoring the decision argument
- Activating every lens or framework
- Rewriting the artifact before exposing what is wrong
- Forcing a specialist scenario when general review is sufficient
- Failing to route when a specialist risk clearly dominates
- Inventing stakeholder beliefs or evidence
- Treating more objections as more rigor
Evaluation Requirements
Evaluate generalist invocation and specialist routing; minimal and rich context; varied artifacts; all modes and entry paths; grounding; prioritization; contradiction detection; framework escape; persuasion; strong-artifact control; non-abusive pressure; and contextual verdicts.
Supporting Files
- Artifact review template
- General artifact logic
- Sources and adaptations
- Bundled runtime contract
- Worked example
- Weak example
Provenance
Derived from PANT's shared antagonist, evidence, concern, feedback, dialogue, and routing contracts. It is intentionally a fallback, not a universal replacement for specialist judgment.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| finishing-a-development-branchobra/superpowersUse when implementation is complete, all tests pass, and you need to decide how to integrate the work | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| systematic-debuggingobra/superpowersUse when encountering any bug, test failure, or unexpected behavior, before proposing fixes | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| test-driven-developmentobra/superpowersUse when implementing any feature or bugfix, before writing implementation code | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| skill-creatoranthropics/skillsCreate new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |