| FMRS | 42 / 100 · D | 79 / 100 · B |
|---|
| Reliability | 7 / 20 | 12 / 20 |
|---|
| Security and permissions | 5 / 20 | 18 / 20 |
|---|
| Maintenance | 9 / 20 | 16 / 20 |
|---|
| Documentation | 11 / 20 | 18 / 20 |
|---|
| Setup experience | 10 / 20 | 15 / 20 |
|---|
| Best for | - Developers familiar with Claude Code, the Claude Agent SDK, or multi-agent workflows
- Teams needing self-learning routing, swarm coordination, and project-level code analysis
- Users building agent applications with the TypeScript API
| - Engineering teams scaling AI coding agents while controlling spend
- Platform teams that need auditable, rollback-aware agent execution records
- Organizations using multiple agents (Claude, Codex, Gemini) under one governed flow
|
|---|
| Not for | - Users who only need one simple MCP tool
- Environments that do not permit project-level command execution or codebase access
- Production environments requiring a verified stable release rather than an alpha version
| - Developers who just want a simple code-completion helper without run governance
- Teams unwilling to accept local JSONL run records with HMAC signing
- Environments without a supported adapter (Claude/Codex/Gemini CLIs, direct provider, or verifier-only)
|
|---|
| Required permissions | - Access to project files and file structure for initialization, pretraining, code analysis, and edit-related hooks
- Execution of npm, npx, build, test, and other user-triggered shell commands
- Git history access during pretraining, plus repository or pull-request access when GitHub tools are used
- Permission for the MCP client to start the local npx-configured service
| - Read/write access to the configured repo root (writable scope restrictable via --allow-path/--deny-path)
- Ability to invoke the chosen agent CLI (Claude/Codex/Gemini, etc.) and its underlying model account
- Local disk writes for run records, receipts, and share outputs (share/run-receipt., etc.)
- Execution of user-configured verifier commands (e.g. npm test)
|
|---|
| Risks and side effects | - The README calls the project production-ready, but installation examples and the roadmap still use alpha releases; stability should be verified
- Self-learning memory may store project-related task inputs, outputs, scores, critiques, and token data
- Agents and background workers may perform broad code analysis, command execution, or automation
- Performance, accuracy, and cost-saving figures are project-reported claims and are not independently verified by the supplied material
- Repository metadata lists the license as unknown, while the README claims MIT; verify the repository LICENSE before use
| - Verifier commands are executed; poorly chosen verifier commands carry security risk despite preflight policy checks
- Cost and token figures may be estimates (labeled with provenance) and should not be treated as settled accounting
- Receipt integrity must be verified before use as trustworthy evidence; missing rollback evidence is flagged as EVIDENCE_BOUNDARY
- Policy defaults come from martin.config.yaml and can be overridden by CLI flags; misconfiguration can loosen safety boundaries
|
|---|
| Supported clients | Claude Code | Claude Code, Codex, Gemini CLI, Cursor, VS Code, GitHub Copilot, Continue |
|---|
| Tools | 11 | 0 |
|---|