Pre-Mortem Analyst

**Quick judgment:** This is a strong, well-scoped analytical skill that leverages a pre-mortem framing to elicit deeper, more honest risk identification than typical risk assessments. It is highly compatible with LLM reasoning capabilities and offers a clear, structured output format that will be easy for users to consume. **Strengths & functionality:** - Encourages a specific cognitive frame (“It has already failed”) that reduces optimism bias. - Provides a stepwise method: set the scene → list 10+ failure reasons → categorize (People/Process/Technology/External) → rate likelihood × impact → define mitigations → define early warning signs. - Output template is explicit and practical, including top-3 priorities and monitoring checklists. - Trigger phrases are broad and intuitive (e.g., “pre-mortem”, “what could go wrong”, “stress test”, “what would kill this”), increasing activation reliability. **Key risks / limitations:** - Quality of analysis will depend heavily on how much concrete context the user supplies about their project; with vague inputs, the model may resort to generic failure modes. - Likelihood × impact ratings are inherently subjective; the skill does not enforce grounding in quantitative data. - If used on harmful or high-risk projects (e.g., dangerous technologies), it could indirectly help with robustness planning unless constrained by higher-level safety filters. - No explicit guidance on handling very large or highly regulated projects where domain expertise and compliance constraints are critical. **Recommended scenarios for use:** - Before launching new products, features, campaigns, or major internal initiatives, to surface non-obvious risks. - As a structured facilitation aid for teams doing risk workshops or design reviews. - In conjunction with strategy or planning skills (e.g., inversion, second-order consequences) to turn identified risks into concrete design or process changes. - As a recurring check (e.g., quarterly) on ongoing projects to reassess emerging failure modes and update warning indicators.