Stage And Risk Selection

Workflow

1. Pick exactly one stage:

- Discovery: explore goals/values/unknowns; avoid fake precision. - Specification: write contract-like requirements; enforce atomic + testable. - Validation: check satisfiability, consistency, gaps, and evidence. - Incident: time-critical mitigation + learning; label uncertainty and act. - Audit/Compliance: focus on prohibitions, access control, and loopholes.

1. Declare risk level (Low/Med/High). If High, require explicit uncertainty handling and a targeted self-audit.
2. Choose the artifact you want:

- Clarifying questions, stakeholder summary, constraints list, requirement set, acceptance criteria, conflict matrix, mitigation plan, etc.

1. Add stage-appropriate guardrails:

- Discovery: allow qualitative terms, require operationalization (examples/rubric/study/TODO metric plan). - Specification: use RFC-2119 modality in Statements; include fit criteria/validation; forbid inventing facts. - Incident: limit questions; provide an actionable plan with confidence labels; interleave need/option/risk. - Audit/Compliance: keep prohibitions as SHALL NOT; run a loophole scan.

1. Add a stop rule:

- “If information is missing, write UNKNOWN (needs follow-up) and list the minimum follow-up questions.”

Prompt snippet (copy/paste)

Stage: <Discovery|Specification|Validation|Incident|Audit/Compliance>
Risk: <Low|Med|High>

Role: Senior requirements engineer.
Rules:
- Do not invent facts; mark missing info as UNKNOWN (needs follow-up).
- Apply stage-appropriate guardrails.

Output: <describe artifact + required schema>
Inputs: <paste the full input text here>