$ AI agent governance, security tooling, and mechanical enforcement.

rigscore v0.2.0

rigscore, security, ai, devtools, mcp, open-source

Changes: v0.1.0 → v0.2.0

Three additions. The check count stays at 8 but the depth increases significantly.

FeatureWhat it does
Cross-config coherenceCompares governance claims against actual configuration — flags contradictions
Deep source secretsRecursive scanning with 34 secret patterns via --deep
MCP supply chainVersion pinning checks and typosquatting detection for MCP server packages

Coherence checking

This is the check that makes rigscore different from generic secret scanners. A CLAUDE.md file might say “no shell access” while the MCP config grants a shell server full filesystem access. A governance file might claim “all secrets encrypted” while .env files sit unprotected.

The coherence check reads your governance claims and cross-references them against your actual configuration. When they contradict, rigscore flags it and applies a compound risk penalty to the overall score.

Configurations that are honestly insecure score better than configurations that lie about their security posture.


Deep source scanning

The --deep flag enables recursive source file scanning across 34 secret patterns — AWS keys, GitHub tokens, database connection strings, private keys, Slack webhooks, and more. Without --deep, rigscore only checks common config file locations.

1
npx rigscore --deep

MCP supply chain

MCP servers are npm packages that your AI agent loads and executes. rigscore now checks whether those packages use pinned versions (not latest or *) and scans package names for typosquatting patterns against known legitimate MCP servers.


Install

1
npx rigscore

MIT licensed. github.com/Back-Road-Creative/rigscore

Configuration details reflect a production environment at time of writing. Implementation specifics vary based on tooling versions, platform updates, and organizational requirements. Validate approaches against current documentation before deployment.