Skip to main content

Repo-native by design

StrayMark has no SaaS, no dashboard, no second source of truth. Every artifact — Charters, AILOGs, agent rules, compliance evidence — lives as a versioned file in your git repository, next to the code it governs.

Why this matters​

  • Reasoning travels with the code. A git blame on a tricky line leads to the commit, which leads to the AILOG, which leads to the Charter, which leads to the risks declared up-front. No tool to log into, no broken link from a moved Confluence page.
  • Diff = governance. When STRAYMARK.md changes, it's a PR with reviewers and a discussion. When an AILOG is missing for a non-trivial commit, CI catches it. The same git workflow you already use is the only workflow.
  • No vendor lock-in. The framework is plain Markdown plus a Rust CLI. Both are MIT-licensed and self-hosted. Walking away from StrayMark is rm -rf .straymark/ — and your history doesn't go with it.
  • Forks are first-class. Organizations can fork the framework, pin a version, and evolve their own rules. The CLI's update-framework command respects that pinning.

How it works​

your-repo/
├── STRAYMARK.md                  # Unified rules — first thing agents read
├── .straymark/
│   ├── 00-governance/            # AGENT-RULES, doc policies, quick reference
│   ├── 01-charters/              # CHARTER-NN-slug.md
│   ├── 02-ailogs/                # AILOG-YYYYMMDD-NN-slug.md
│   ├── 03-decisions/             # AIDEC, ADR
│   ├── 04-models/                # MCARD
│   ├── 05-security/              # SEC, ETH, DPIA
│   ├── 06-evolution/             # TDE, deprecations
│   └── audits/CHARTER-NN/        # External audit reports
└── src/...                       # Your actual code

The CLI (straymark validate, straymark audit, straymark compliance) operates on the filesystem, deterministically. No daemons, no background sync.

Learn more​