Decision Critic
The Decision Critic analyzes choices and provides insightful feedback on potential outcomes, aiding users in making more informed decisions.
Install on your platform
We auto-selected Claude Code based on this skill’s supported platforms.
Run in terminal (recommended)
claude mcp add decision-critic npx -- -y @trustedskills/decision-critic
Or manually add to ~/.claude/settings.json
{
"mcpServers": {
"decision-critic": {
"command": "npx",
"args": [
"-y",
"@trustedskills/decision-critic"
]
}
}
}Requires Claude Code (claude CLI). Run claude --version to verify your install.
About This Skill
What it does
The decision-critic skill provides feedback on the quality of decisions made by an AI agent. It analyzes proposed actions and offers suggestions to improve their reasoning, alignment with goals, and overall effectiveness. This helps refine the agent's decision-making process and leads to better outcomes in complex scenarios.
When to use it
- Improving task planning: Use it when you want your agent to generate more robust and well-reasoned plans for achieving a goal.
- Refining conversational agents: Apply it to assess the appropriateness and helpfulness of responses in dialogue, ensuring they are aligned with user intent.
- Evaluating complex decision chains: Employ it to review sequences of actions where each step builds on the previous one, identifying potential pitfalls or suboptimal choices.
- Debugging agent behavior: Leverage it when an agent is exhibiting unexpected or undesirable behaviors and you need to understand why it made certain decisions.
Key capabilities
- Provides feedback on decision quality
- Suggests improvements to reasoning
- Evaluates alignment with goals
- Identifies potential issues in action sequences
Example prompts
- "Critique the following plan: [agent's plan]"
- "Evaluate this proposed response: [agent's response] and suggest alternatives."
- "I want to achieve [goal]. The agent decided to [action]. Is that a good idea? Why or why not?"
Tips & gotchas
The effectiveness of decision-critic depends on the clarity and detail provided in the agent’s initial decision. Ensure the context and reasoning behind the action are well-defined for optimal feedback.
Tags
TrustedSkills Verification
Unlike other registries that point to live repositories, TrustedSkills pins every skill to a verified commit hash. This protects you from malicious updates — what you install today is exactly what was reviewed and verified.
Security Audits
| Gen Agent Trust Hub | Pass |
| Socket | Pass |
| Snyk | Pass |
🌐 Community
Passed automated security scans.