Skip to main content
Glama
Keesan12

MartinLoop — Stop Runaway Loops, Bad Code & Token Waste

by Keesan12
OverviewSchemaSourceRelated ServersDiscussions(1)Admin
TypeScript
Local

Related Servers

Alternatives to MartinLoop — Stop Runaway Loops, Bad Code & Token Waste

Related Servers

  • Agent Cost MCP

    vk0dev
    A
    license
    Not graded
    quality
    A
    maintenance
    Local Cost Guard and runtime guardrails for AI agents in Claude Code.
    126 npm
    MIT
  • agentguard by agentwares

    agentguardofficial

    agentwares
    A
    license
    Not graded
    quality
    C
    maintenance
    Enforces policy controls for AI agents, including spend limits, action approvals, kill switch, scoped credentials, dry-run diffs, loop prevention, and auditable hash-chained logs.
    MIT

  • TokenControlPlane

    devthinker-ai
    A
    license
    Not graded
    quality
    B
    maintenance
    Stop runaway agents. One gateway in front of all your agent traffic — MCP tools and LLM models — one budget, one kill switch. Sits between AI clients (Cursor, Claude Desktop, Windsurf) and remote MCP servers / OpenAI-compatible model APIs; hard token budgets; a single static Go binary.
    1
    Apache 2.0

  • dashclaw

    ucsandman
    A
    license
    B
    quality
    A
    maintenance
    Governance runtime for AI agents: a guard tool evaluates risky actions against policy before they execute (block / warn / require human approval), approvals route to a human queue, and every action becomes a replayable decision record with per-action spend tracking. Runs over stdio via npx @dashclaw/mcp-server; works with Claude Code, Codex, LangChain, CrewAI, or any MCP host.
    68
    306
    MIT

  • agentos-mcp-server

    imran-siddique
    A
    license
    Not graded
    quality
    F
    maintenance
    Governance kernel for AI agents — policy enforcement, code safety verification, multi-model hallucination detection (CMVK), trust attestation (IATP), and immutable audit trails. Works with Claude Desktop, Cursor, and any MCP client.
    74
    MIT
Deploy Server
One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

TDQS

A3.7/5.0

Scored across 24 tools

Disambiguation3/5

Several tools have overlapping purposes: martin_get_run, martin_status, martin_run_dossier, martin_dossier, and martin_inspect all read run data with subtle differences in format and depth. The descriptions try to disambiguate with 'do not use' guidance, but an agent could easily confuse martin_dossier vs martin_run_dossier vs martin_get_run.

Naming Consistency4/5

Most tools follow a consistent martin_<verb>_<noun> pattern (martin_get_run, martin_create_pr, martin_review_pr). Minor deviations exist: martin_doctor, martin_arcade, and martin_arcade_status use noun-only or different structures, and martin_status vs martin_get_status inconsistency appears in descriptions.

Tool Count3/5

24 tools is at the high end of acceptable for a complex run-management system, but the count feels inflated by near-duplicate read surfaces (dossier, run_dossier, get_run, inspect, status, logs, verification_results). A leaner set of 15-18 tools would cover the same functionality.

Completeness4/5

The lifecycle is well covered: estimate, plan, preflight, run, pause/cancel/continue, status, verification, eval, PR creation/review. Minor gaps include no explicit tool for listing attempts across runs or deleting/archiving runs, but agents can work around these.

Maintenance

ActivityActive
ResponsivenessSlow

Resources