# agent-readiness-mcp [Health: Active]

**Category:** 💻 Developer Tools  
**Repository:** https://github.com/forgemeshlabs/agent-readiness-mcp  
**GitHub Stars:** 0  
**npm Downloads (last month):** 1176  
**Views:** 0  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/agent-readiness-mcp

## Description
Agent Readiness MCP — the free Agent Readiness scanner as an MCP server. Measures your ASO Score.

## Tools
Capabilities this server exposes over MCP:

- **scan_site** — Scan a website for Agent Readiness using the ASO (Agent Signal Optimization) framework and return an Agent Readiness Report. Use this for a full site-level baseline, competitive audit, or before/after readiness measurement; use check_signal instead when you only need one named signal, and use get_fix_plan when you only need remediation steps. Runs 34 checks across discoverability (robots.txt, sitemap, llms.txt, DNS-AID, Link headers), content accessibility (markdown negotiation), bot access control (AI bot rules, Content Signals, Web Bot Auth), invocation (API catalog, OAuth discovery, OAuth protected resource, auth.md, MCP Server Card, Google A2A Agent Card, Agent Skills, WebMCP), commerce (x402, MPP, UCP, ACP, pricing), Google generative AI search basics, browser-agent UX, and identity/trust signals. Returns the ASO Score (0-100, formally the Agent Readiness Index), ASO maturity level (ASO-0 Invisible … ASO-5 Autonomous-Commerce-Ready), an agent-readiness verdict, per-pillar scores, per-check evidence, and prioritized recommendations.
- **check_signal** — Run one specific agent-readiness check against a site (e.g. 'a2a-agent-card', 'llms-txt', 'mcp-server-card', 'x402'). Use this for targeted validation after making a fix or when debugging one signal; use scan_site for the complete ASO Score and get_fix_plan for a prioritized remediation roadmap. Use list_checks first when you need valid check ids. Returns status, evidence, and a fix recommendation, and omits raw remote artifacts by default.
- **list_checks** — List the full catalog of supported ASO checks with id, name, category, description, and spec link. Use this before check_signal to discover valid check ids, to build UI filters, or to explain the scanner coverage; it does not scan a site or produce a score.
- **get_fix_plan** — Scan a site and return a prioritized remediation plan: which signals to add first, the ASO Score points each fix is worth, and ready-to-paste artifact templates (robots.txt AI rules, llms.txt, agent.json, A2A agent-card.json, MCP server card, x402 manifest, pricing.json, security.txt, status endpoint). Use this when the user wants an implementation roadmap or copy-paste fixes; use scan_site when they need full evidence and per-check scoring, and use check_signal to verify one completed fix.
- **get_aso_framework** — Return the ASO (Agent Signal Optimization) framework reference: the six signal pillars with point weights, the Agent Readiness Index maturity levels (ASO-0 through ASO-5), certification thresholds, and the scoring rubric. Use this for education, documentation, or explaining how scores are calculated; it does not fetch or scan a website. Source: https://agentsignaloptimization.com

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `npx` (confidence: high):

```json
"mcpServers": {
  "agent-readiness-mcp": {
    "command": "npx",
    "args": ["-y","@forgemeshlabs/agent-readiness-mcp"]
  }
}
```

## Documentation

## What agent-readiness-mcp does

The agent-readiness-mcp MCP server evaluates a website against the Agent Signal Optimization (ASO) framework. A full scan produces an Agent Readiness Report containing an overall score from 0 to 100, a maturity level, scores for six pillars, and results for 34 individual checks. Each check is reported as pass, partial, or fail with supporting evidence and a recommendation.

The checks cover discovery, content, bot access, interoperability, commerce, identity, and trust. Examples include robots.txt, sitemaps, Link headers, DNS-AID, AI-related well-known files, Markdown content negotiation, llms.txt, crawler rules, API and OAuth discovery, MCP Server Cards, Google A2A Agent Cards, Agent Skills, WebMCP, x402, MPP, UCP, ACP, machine-readable pricing, HTTPS, JSON-LD, OpenAPI, agent.json, security.txt, status endpoints, versioning, and cross-file consistency.

The score rolls up into six weighted pillars: Discoverability, Identity, Trust, Commerce, Reputation, and Memory. Maturity levels range from ASO-0 Invisible through ASO-5 Autonomous-Commerce-Ready. The results are directional self-assessments; certification requires separately verified evidence.

## How it works

The MCP tools accept a website target for scanning or signal inspection. `scan_site` runs the complete assessment, while `check_signal` evaluates one named signal such as `llms-txt`, `a2a-agent-card`, or `x402`. `list_checks` helps identify available signal IDs and links them to relevant specifications. `get_aso_framework` explains the pillars, weights, levels, and certification thresholds without scanning a site.

`get_fix_plan` converts scan findings into a prioritized remediation sequence. Its output can include ready-to-paste starting points for AI crawler rules, llms.txt, agent.json, A2A and MCP cards, x402 manifests, pricing.json, security.txt, and status endpoints. The templates address the findings; they do not themselves deploy changes to the target website.

## Setup and configuration

The agent-readiness-mcp MCP server requires Node.js 18 or newer. Install the published `@forgemeshlabs/agent-readiness-mcp` package globally, or run it directly with `npx`, which is the documented approach for MCP clients. The package uses local stdio transport and does not require API keys, tokens, cookies, or third-party credentials.

Claude Code can register the server with `claude mcp add aso -- npx -y @forgemeshlabs/agent-readiness-mcp`. Claude Desktop, Cursor, Windsurf, and other MCP clients can configure a server entry whose command is `npx` and whose argument is `-y @forgemeshlabs/agent-readiness-mcp`. Building from source is intended for scanner development and requires cloning the repository, installing dependencies, and running the build.

## Tools and capabilities

- `scan_site`: Run all 34 checks and return the score, level, pillar results, evidence, and recommendations.
- `get_fix_plan`: Produce a prioritized remediation plan with templates for supported agent-facing files and endpoints.
- `check_signal`: Test one named signal for a target website.
- `list_checks`: Show the scanner catalog and specification links.
- `get_aso_framework`: Return the scoring rubric and maturity model.

## Limitations and notes

The agent-readiness-mcp MCP server is beta software, and its ASO scoring may change as agent standards develop. Keeping llms.txt in the scan reflects use by some non-Google agents; the documentation states that Google Search itself ignores that file. The scanner also includes browser-agent checks for semantic controls, labels, ARIA or role fallbacks, and hidden overlays.

Scanning makes outbound requests to URLs supplied by the user. The implementation only permits HTTP and HTTPS, blocks private and reserved destinations after DNS resolution, pins requests to the validated public IP, validates redirects individually, and limits redirect hops to five. The npm package does not enable public HTTP deployment; any public deployment needs authentication, per-client rate limits, request logging, and an egress policy.

The server does not modify scanned sites. Its output is an assessment and, where available, remediation guidance. Use the report as a baseline for prioritizing work rather than as proof of certification.

_Full upstream README: https://allmcps.com/mcp/agent-readiness-mcp/readme_