AI agent governance, security tooling, and mechanical enforcement.

Headless Mode

rigscore scores your AI dev config. 28 checks — 13 scored plus 15 advisory. A number out of 100 and a CI-ready exit code. Runs locally, no account.

Install & run
npx github:Back-Road-Creative/rigscore
  ╭────────────────────────────────────╮
  │                                    │
  │        HYGIENE SCORE: 78/100       │
  │        Grade: B                    │
  │        Risk: Standard              │
  │        Practice: 85/100 (B)        │
  │                                    │
  ╰────────────────────────────────────╯
Full sample →
MCP supply chain

Pins your MCP server configs and flags when they change. Catches typosquats, version drift, and config swaps.

Coherence

Reads your governance file and checks whether the rest of your config backs up the claims it makes.

Local only

No account, no network. --online is opt-in when you want a registry lookup.

Behavioral rules for AI agents are text in the context window. Under pressure — deep in a fix loop, resolving conflicting instructions, running low on context — the model rationalizes around them. This isn’t a hypothetical failure mode. It’s documented.
Latest release rigscore v2.1.0 2026-07-15