mcpField
M
MCP Agent Health Reporter

MCP Agent Health Reporter

Local
raw.githubusercontent.comDeveloper Tools0Updated 15w ago
RepositoryVisit website

About MCP Agent Health Reporter

MCP Agent Health Reporter scans agents under AOS discipline and reports readiness scores. It is intended for teams checking agent health before deployment, with both single-agent scans and directory-wide audits.

Key features

  1. 01Scan an agent and report its score
  2. 02Audit agents across a directory
  3. 03Report certification status
  4. 04Record calls as JSONL evidence

Use cases

  1. 01Checking an agent before deployment
  2. 02Auditing a directory of agents
  3. 03Reviewing readiness scores and certification

MCP support

It exposes the `aos_agent_health_scan` and `aos_agent_health_audit` tools to MCP clients. The server uses stdio by default.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

mcp.json · Universal config
{
  "mcpServers": {
    "aos-standard-mcp-agent-health": {
      "command": "uvx",
      "args": [
        "mcp-agent-health"
      ]
    }
  }
}
Run commanduvx mcp-agent-health

Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.

Packages
pypimcp-agent-health0.3.1

Tools

No tool list yet

Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.

Details

From the project README

MCP Agent Health Reporter

Renamed (2026-07): Formerly aos-agent-health-mcp / CLI aos-health-mcp. Migrate with pip install mcp-agent-health==0.3.1 and update your MCP config command to mcp-agent-health.

MCP server that validates agent health under AOS discipline and reports readiness scores. Use it to catch agents that silently fail in production — healthy locally, broken when deployed.

Install

pip install mcp-agent-health==0.3.1
# or editable from source:
# pip install -e ./packaging

MCP Client Configuration

Add to your mcp.json (Cursor / Claude Desktop):

{
  "mcpServers": {
    "mcp-agent-health": {
      "command": "mcp-agent-health",
      "args": [],
      "env": {
        "AOS_HEALTH_TARGET_DIR": "/path/to/your/agents"
      }
    }
  }
}

Tool: aos_agent_health_scan

Request (non-default args — real signal):

{
  "tool_id": "1066",
  "target_dir": "/home/builder/my-agents",
  "mock": false
}

Response:

{
  "status": "success",
  "aos_score": 62.5,
  "certified": false,
  "tool_id": "1066",
  "target_dir": "/home/builder/my-agents",
  "target_path": "/home/builder/my-agents/1066_AOS_Agent_Health_Reporter",
  "sections": {
    "manifest_declared": 12.5,
    "systemd_runtime": 0.0,
    "immune_loop": 0.0,
    "physical_evidence": 25.0
  },
  "precedent_meta": {
    "precedent_id": "f47ac10b-58cc-4372-a567-0e02b2c3d479",
    "tool_id": "1066",
    "created_date": "2026-06-14T12:00:00+00:00",
    "source_signal_hash": "a1b2c3d4"
  },
  "scanned_at": "2026-06-14T12:00:00+00:00"
}

Tool: aos_agent_health_audit

Request:

{
  "target_dir": "/home/builder/my-agents",
  "mock": false
}

Response (excerpt):

{
  "status": "success",
  "total_tools": 3,
  "avg_score": 54.2,
  "target_dir": "/home/builder/my-agents",
  "reports": []
}

Self-Test (G1 evidence)

python3 packaging/scripts/smoke_self_call.py
# Appends one line to ~/.aos_health_mcp/calls.jsonl

Canonical source sync (TASK B)

正本: core/ahr_scan_pure.py（直接編集はこちらのみ）
配布コピー: packaging/aos_health_mcp/_scan_pure.py（直接編集禁止 — 同期スクリプト経由）

python3 packaging/scripts/sync_scan_pure.py          # copy + verify
python3 packaging/scripts/sync_scan_pure.py --check-only  # CI gate

一致検証: pytest tests/test_logic.py::TestAosAgentHealthReporter::test_scan_pure_canonical_sync

G2 clean install (mock=false)

python3 -m venv /tmp/aos-g2-venv
/tmp/aos-g2-venv/bin/pip install -e packaging/
/tmp/aos-g2-venv/bin/python packaging/scripts/g2_clean_install_smoke.py

Environment Variables

VariableDescription
AOS_HEALTH_TARGET_DIRDefault scan root (fallback: cwd)
AOS_HEALTH_MCP_LOGJSONL log path (default: ~/.aos_health_mcp/calls.jsonl)
AOS_HEALTH_MCP_TRANSPORTstdio (default)

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.