Track falsifiable predictions and compare them against real outcomes over time.
Copy the install command and let the AI configure it · recommended for beginners
Please install the "io.github.commet/argus-decision-mcp" MCP server from askskill: Run: claude mcp add 'io-github-commet-argus-decision-mcp' -- npx -y argus-decision-mcp
Save a falsifiable prediction: within two weeks after the new feature launches, 7-day retention will increase by 3 percentage points. Prepare fields for recording the actual outcome later.
A structured prediction entry with the claim and a place to record the eventual outcome.
Record this prediction and later add the real result next month: after this month's campaign, cost per registration will decrease by 10%.
A trackable prediction record that can later be updated with actual data to verify accuracy.
Save a falsifiable hypothesis: under the same conditions, method A will outperform method B by at least 2% accuracy. I want to record the real result after the experiment.
A verifiable hypothesis record that supports post-experiment comparison with observed results.
Product managers, researchers, or analysts can turn a judgment into a falsifiable prediction before acting, then record what actually happened afterward to reduce hindsight bias.
When a team wants to review whether a forecast was accurate, this tool pairs the prediction with the real outcome so evaluation is grounded in facts rather than impressions.
It saves falsifiable predictions and later records what actually happened in reality. Its core idea is that reality grades the prediction, not the model.
It suits situations where you make judgments and later review them, such as product decisions, research hypotheses, or tracking forecasts with data. It emphasizes verifiability and comparison with outcomes.
The provided material does not include installation steps, runtime requirements, or key requirements. See the source repository for details.
Run end-to-end browser testing and quality debugging directly from your IDE.
Browse Metaculus questions and submit forecasts and rationale via API tokens.
Assess evidence behind org design and leadership claims, and suggest safer wording.
Deterministically verify model outputs for evidence, contradictions, calibration, and provenance.
Log and retrieve decisions with provenance, supersession, and audit trails.
Make structured decisions with AHP, pairwise comparisons, rankings, and consistency checks.