AymanShams·AymanShams/codex-coding-os/.agents/skills/pre-mortem/SKILL.md

pre-mortem

Skill1 starsChanged 2 months agoMPL-2.0View on GitHub

Use when the user asks to run a premortem or failure-first risk analysis on a concrete plan, launch, product, pricing change, hire, partnership, strategy, SOP, operating model, implementation, project, rollout, or high-cost decision before or during execution. Use for requests like stress test this plan, what could kill this, what could go wrong, find blind spots, what am I missing, where will this break, future-proof this, poke holes in this plan, identify failure modes, pressure-test before launch, reduce execution risk, or revise the plan to avoid failure. The skill assumes the initiative has failed in the future, works backward to find real causes, then produces hidden assumptions, warning signs, mitigation actions, and a revised plan. Do not use for simple factual questions, vague ideas without a concrete plan, ordinary critique of an existing argument, lightweight feedback, creative editing, grammar review, or decisions that are already irreversible. If the primary task is skeptical review of an artifact or recommendation, use deep-critic. If the primary task is reviewing an SSOT-style SOP, policy, workflow, governance document, or controlled artifact, use ssot-auditor. If the primary task is acceptance criteria, readiness checklist, defect log, evidence requirements, or pass/fail approval, use artifact-validation-workflow. If the primary task is building a delivery plan, dependency tree, roadmap, or WBS, use wbs-artifact-planner.