Pre-Mortem Analyst

Imagine the project already failed, then work backward to find why. More powerful than risk assessment because it assumes failure is certain. Use when user says "pre-mortem", "premortem", "imagine this failed", "what could go wrong", "risk analysis", "before we launch", "stress test", "what would kill this", "project risks".

Security Scan

This is an instruction-only pre-mortem facilitation skill whose required artifacts and runtime instructions are consistent with its description and do not request credentials, installs, or system access.

Purpose & Capability

The name/description (pre-mortem analysis) matches the contents: facilitation steps, templates, examples, and frameworks for imagining failure. There are no unrelated requirements (no env vars, binaries, or installs) that would be unexpected for this purpose.

Instruction Scope

SKILL.md and reference docs provide step-by-step facilitation guidance, output templates, examples, and monitoring suggestions. The instructions do not tell the agent to read files, access environment variables, contact external endpoints, or collect system data beyond the user-provided project context.

Install Mechanism

No install spec or code files that would be written to disk or executed are present. Being instruction-only minimizes supply-chain/code-execution risk.

Credentials

The skill requests no environment variables, credentials, or config paths. This is proportionate for a facilitation/analysis skill that only needs user-provided project context.

Persistence & Privilege

always is false and default autonomous invocation is unchanged. The skill does not request persistent presence, system modification, or access to other skills' config.

Assessment

This skill is instruction-only and internally consistent with being a pre-mortem facilitator. It does not request credentials or install code, so supply-chain and credential-exfiltration risks are low. However, be cautious about the project data you paste into any third-party model or agent — do not include secrets, credentials, or sensitive PII in prompts. If you plan to run pre-mortems on proprietary projects, confirm how your agent/provider handles and stores prompt content and consider running initial tests with redacted or synthetic data. Lastly, review the examples/templates to ensure the output format and level of detail match your team's needs before using it in a live workshop.

Why Pre-Mortem > Risk Assessment

Risk Assessment: "What MIGHT go wrong?" → Optimism bias filters answers Pre-Mortem: "It's 6 months later. It FAILED. Why?" → Liberates honest analysis

Research: Pre-mortems increase problem identification by 30%.

The Process

1. Set the scene: "It's [date]. This has failed completely."
2. Brainstorm causes: List 10+ failure reasons (no filtering)
3. Categorize: People, Process, Technology, External
4. Rate: Likelihood × Impact (H/M/L)
5. Prevent: Top 3 get specific mitigation actions
6. Monitor: Define early warning signs

Output Format

PROJECT: [Name]
FAILURE SCENARIO: "It's [date]. [Project] has completely failed."

WHY IT FAILED:

👥 PEOPLE: [Cause] - L×I: H/H | Prevent: [x] | Warning: [y]
⚙️ PROCESS: [Cause] - L×I: M/H | Prevent: [x] | Warning: [y]
💻 TECHNOLOGY: [Cause] - L×I: L/H | Prevent: [x] | Warning: [y]
🌍 EXTERNAL: [Cause] - L×I: M/M | Prevent: [x] | Warning: [y]

TOP 3 PRIORITIES:
1. [Risk] → [Specific action]
2. [Risk] → [Specific action]
3. [Risk] → [Specific action]

WARNING SIGNS TO MONITOR:
□ [Early indicator 1]
□ [Early indicator 2]

Common Failure Categories

Category	Common Causes
People	Key person leaves, skill gaps, misalignment, low buy-in
Process	Aggressive timeline, scope creep, dependency issues
Tech	Doesn't scale, integration fails, security breach
External	Market shift, competitor move, regulation change

Integration

Compounds with:

• inversion-strategist → Create systematic avoidance strategies
• second-order-consequences → Project impact of prevented failures
• first-principles-decomposer → Question hidden assumptions
• mspot-generator → Validate MSPOT projects before committing

---

See references/examples.md for Artem-specific pre-mortems