| FMRS | 34 / 100 · D | 79 / 100 · B |
|---|
| Reliability | 3 / 20 | 12 / 20 |
|---|
| Security and permissions | 6 / 20 | 18 / 20 |
|---|
| Maintenance | 13 / 20 | 16 / 20 |
|---|
| Documentation | 8 / 20 | 18 / 20 |
|---|
| Setup experience | 4 / 20 | 15 / 20 |
|---|
| Best for | - AI architects needing persistent memory and cross-model coordination
- Engineers building enterprise multi-window applications, trading platforms, or IDE-class tools
- Researchers studying self-evolving systems, gated recursive self-improvement, and multi-agent governance
| - 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 | - Static content sites or simple blogs
- Teams seeking a drop-in syntax replacement
- Developers unwilling to adopt the Worker Actor Model or treat AI as a peer maintainer
| - 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 source code, documentation, issues, pull requests, and discussions for knowledge-base construction
- Persistent storage for agent memory, messages, and graph data
- If GitHub Workflow is enabled, management of issues, pull requests, reviews, labels, and projects
- If Neural Link is enabled, inspection of live application state and mutation of the running application
- Sandboxed file access may be used by internal Neo.ai.Agent local loops
| - 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 | - Neural Link can modify live application state and hot-patch code, so incorrect operations may affect running applications
- GitHub workflows may create or review pull requests, manage issues, and synchronize repository state
- Persistent memory and knowledge stores may retain repository content, discussions, and agent reasoning
- Multi-tenant deployments must preserve tenant identity and visibility isolation; the README gives no specific security configuration
| - 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, Codex, Gemini CLI, Cursor, VS Code, GitHub Copilot, Continue |
|---|
| Tools | 0 | 0 |
|---|