30
GitHub Stars
1
Bundled Files
3 weeks ago
Catalog Refreshed
2 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstart where the catalogue uses aiagentskills.
npx veilstart add skill lyndonkl/claude --skill deliberation-debate-red-teaming- SKILL.md10.9 KB
Overview
This skill runs structured adversarial reviews (deliberation, debate, and red teaming) to expose blind spots, hidden assumptions, and vulnerabilities in plans or decisions before launch. It guides role assignment, critique generation, risk scoring, and concrete mitigation recommendations so teams can make safer, better-informed choices.
How this skill works
You provide a specific proposal and stakes; the skill assigns 3–5 adversarial roles (e.g., security, legal, operations, competitor, customer) and generates realistic critiques from each perspective. It synthesizes findings, scores risks by severity and likelihood, and produces prioritized mitigations such as design changes, safeguards, reduced blast radius, contingency plans, or monitored acceptance.
When to use it
- Before launching a product feature or major change to identify showstoppers
- When you want adversarial review to counter groupthink or confirmation bias
- To stress-test strategy against worst-case or competitor scenarios
- When stakeholders ask to "play devil's advocate" or "what could go wrong"
- To validate legal, security, or operational readiness prior to release
Best practices
- Scope crisply: evaluate a specific proposal, timeline, and reversal options
- Choose 3–5 roles tailored to the proposal—mix external, internal, and devil's-advocate viewpoints
- Be adversarial but realistic: steelman critiques rather than creating strawmen
- Score each risk by severity × likelihood and classify as showstopper/high/monitor/accept
- Map each priority risk to a concrete mitigation with owner, timeline, and verification criteria
Example use cases
- Assessing readiness to roll out a new feature to all users, identifying security, legal, and ops gaps
- Validating a go-to-market strategy by simulating competitor and regulator responses
- Stress-testing an infrastructure migration for runbooks, rollback plans, and blast-radius limits
- Running a pre-mortem on a funding or partnership decision to surface hidden liabilities
- Challenging assumptions in a product roadmap to prioritize high-impact fixes before release
FAQ
Pick 3–5 roles that are most likely to reveal relevant blind spots; too many dilutes focus, too few misses perspectives.
When is red teaming not appropriate?
Skip it for trivial, low-stakes decisions, urgent emergencies that require immediate action, or when the proposal is already thoroughly vetted.