| FMRS | 63 / 100 · C | 79 / 100 · B |
|---|
| Reliability | 10 / 20 | 12 / 20 |
|---|
| Security and permissions | 12 / 20 | 18 / 20 |
|---|
| Maintenance | 13 / 20 | 16 / 20 |
|---|
| Documentation | 14 / 20 | 18 / 20 |
|---|
| Setup experience | 14 / 20 | 15 / 20 |
|---|
| Best for | - Developers and teams building AI agents that need extra oversight to prevent reasoning lock-in and over-engineering.
- Users working with MCP clients like Claude Desktop, Cursor, Windsurf, or VS Code who want long-horizon workflows to stay on track.
- Developers seeking research-backed reliability improvements for agents.
| - 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 | - Simple or fast tasks where an external oversight call adds unwanted latency and cost.
- Scenarios where you cannot or do not want to rely on external LLM API calls (requires at least one provider key).
- Projects requiring active feature development or official support, as the server is in maintenance mode.
| - 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 | - Requires at least one LLM provider API key (Gemini, OpenAI, OpenRouter, or Anthropic) to function.
- HTTP mode opens a network port; recommend restricting to loopback for security.
- May modify client configuration files (e.g., Claude Desktop, Cursor, Windsurf, VS Code) during installation to add the server entry.
| - 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 | - Sending prompts and context to external LLM APIs; ensure sensitive data is not exposed.
- HTTP mode with non-loopback access increases attack surface; v2.9.0 adds host validation and CORS restrictions but misconfiguration could cause outages.
- Over-frequent `vibe_check` calls can slow down agent execution and increase costs; optimal dose is ~10-20% of steps.
| - 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 Desktop, Cursor, Windsurf, VS Code | Claude Code, Codex, Gemini CLI, Cursor, VS Code, GitHub Copilot, Continue |
|---|
| Tools | 5 | 0 |
|---|