Evaluate reasoning steps, assumptions, and biases before presenting conclusions.
npx skills add https://github.com/iAMv1/agent-os --skill meta-cognition
Please help me install this Agent Skill. Skill: meta-cognition Source: https://github.com/iAMv1/agent-os/tree/main/skills/meta-cognition Command: npx skills add https://github.com/iAMv1/agent-os --skill meta-cognition
Meta-cognition helps AI agents and humans improve decision quality by evaluating thinking processes, identifying biases, and verifying conclusions before presenting results.
Prompt the agent to perform a structured self-review of its reasoning before delivering an answer.
High-intent search queries and answers about installing and using this skill.
To improve reasoning reliability, you can prompt the AI to perform a structured self-evaluation, tracing its reasoning steps, listing assumptions, and calibrating confidence before delivering an answer.
Checking for biases involves applying a self-evaluation process where the AI explicitly reviews its reasoning, steel-mans the opposition, and verifies conclusions before presenting the final results.
The best way to trace assumptions is to require explicit steps for documenting them, calibrating confidence levels, and performing a post-analysis review to ensure thinking quality in high-stakes scenarios.
Yes, you can use self-evaluation during post-hoc reviews to identify biases, verify conclusions, and explicitly document lessons learned for continuous improvement of future decision-making.
Critical-thinking checks should be applied to high-stakes or complex problems where thinking quality matters, specifically by verifying conclusions and evaluating the underlying reasoning process.
Calibrating confidence helps traceability by forcing the explicit documentation of reasoning steps and assumptions, making it easier to evaluate decision quality and review the thinking process.
intelligentcode-aiDecompose complex decisions into explicit step-by-step analysis and validation plans.
haabeDocument lessons learned from completed delivery increments into memory artifacts.
iamasters-academyConduct structured pre-mortem analysis and steel-man argumentation for project planning.
dzackgarzaIdentify and mitigate LLM cognitive failure modes in high-stakes reasoning.
