Documentation
Everything you need to understand and use ShipSleuth.
ShipSleuth turns fragmented public GitHub activity into an honest diligence read. These docs cover what it measures, how to integrate it, and how to use it responsibly.
Methodology
How ShipSleuth reads public GitHub activity through direct metrics, scoring dimensions, and built-in caveats.
- Author commits, merged PRs, releases, contributor breadth, active days
- Scoring: volume, consistency, breadth, collaboration, releases, recency, concentration
- Percentile anchors derived from real GH Archive data (~8.2M active human accounts)
- Full "View the Math" transparency: see anchors, interpolation, and population estimates per metric
API reference
Three JSON endpoints for single-target analysis, multi-target ranking, and historical snapshots.
- POST /api/analyze — single-target diligence
- POST /api/rank — multi-target ranking
- POST /api/history — adjacent time-window snapshots
- All endpoints return structured JSON with caveats
MCP integration
Connect ShipSleuth to any MCP-compatible AI agent for programmatic GitHub diligence.
- analyze_target — structured public GitHub diligence
- compare_targets — head-to-head comparison
- rank_targets — multi-target ranking
- list_radar_packs — discover curated presets
Responsible use
What ShipSleuth is good at, what it is not, and how to use it without causing harm.
- Public signal only — not a complete picture
- Designed for diligence context, not hiring decisions
- Caveats are first-class: bot share, concentration, truncation
- Never treat commit counts as a proxy for individual productivity