Decision Architect Free for Openclaw

1. Ensure you are using an AI agent environment that can load SKILL.md-based skills.
2. No API key is required for Openclaw Skills; the skill is prompt-driven and does not make network requests.
3. Install or place the skill in your agent’s supported skill directory if your platform requires local skill files.
4. Invoke the skill in chat or agent mode by providing the decision input, at least two options when comparison is needed, and any relevant constraints.

# Example: verify the skill file is available locally
ls -la skills/decision-architect-free/

# Example: start an agent session that can read SKILL.md-based skills
claude

5. For best results, provide the decision context, the alternatives, and any priorities such as cost, speed, risk, or maintainability.
6. Review the returned framework match, bias scan, and confidence label, then refine the prompt if more context is needed.

Openclaw Skills for Decision Architect Free use a simple input/output schema optimized for conversational decision analysis.

Output taxonomy

• Decision type: strategic, tactical, or operational
• Framework: decision matrix, cost-benefit, OODA-style, or reversibility reasoning
• Option comparison: scored table or structured tradeoff summary
• Bias signals: sunk cost, confirmation bias, loss aversion, availability bias
• Confidence label: 🟢 high, 🟡 medium, or 🔴 low

Behavioral rules

• Strategic and tactical inputs are scanned for cognitive bias.
• Operational inputs may skip bias scanning to reduce noise.
• Single-option inputs are auto-completed with "maintain current state" as a comparison baseline.
• No persistent memory, no stored decision history, and no cross-session learning in the free tier.