Audit decisions with your AI to uncover cognitive biases and challenge assumptions.
Copy the install command and let the AI configure it · recommended for beginners
Please install the "io.github.urbanmorph/irrational" MCP server from askskill: Run: claude mcp add --transport http 'io-github-urbanmorph-irrational' 'https://irrational.pages.dev/mcp/server'
Act as an adversarial bias-audit assistant and review my decision to prioritize feature A over feature B. Using confirmation bias, anchoring, sunk cost, and groupthink, identify flaws in my reasoning and provide a checklist of questions I should validate.
A decision-audit report grouped by bias type, with risks, counterarguments, and validation questions.
Audit my interview conclusion about a candidate: I think they are a great fit because communication was smooth and my first impression was positive. Check for halo effect, similarity bias, overconfidence, or recency bias, and suggest more objective evaluation criteria.
A review highlighting possible interview biases and adding structured evaluation criteria.
Run a bias audit on my current research conclusion: I believe the experiment results are sufficient to support the hypothesis. Challenge it from selective sampling, survivorship bias, outcome bias, and motivated reasoning, and list the additional data and evidence I should examine.
A counter-review checklist for research reasoning that reveals weak evidence and biased inference.
Improve complex reasoning with multi-agent debate, bias detection, and structured thinking.
Log and evaluate AI decisions against authority context with clear risk outcomes.
Review system designs from six critical perspectives and produce unified risk assessments.
Safely let your AI run code, query databases, use LMs, and commit to GitHub.
Capture rendered UI, spot design drift, and generate actionable redesign diffs.
Estimate reasoning budgets for AI tasks and visualize effort-time tradeoffs.