about_agentavow
ClaudeWhat AgentAvow is, which tool to use for what, how to read a verdict, and example scans. Call this for an overview or when getting started. No input, read-only.
Developer-Tools
Overview
AgentAvow tells you whether a tool is safe for your AI agent to connect to, before it connects. Point it at a GitHub repo, an npm, PyPI, crates, Docker or Hugging Face package, or a live MCP server. Every result leads with one of three answers and the reason behind it: - Safe to connect: nothing blocking was found. - Review before you connect: a high-severity finding, a published advisory for that version, a deprecated package, or a low trust score. - Do not connect: a critical finding, a known-malicious dependency, or a planted test credential leaving our sandbox. "Certified" marks tools that also passed every provenance check. Under the answer sit two scores: a 0 to 100 trust score built from the findings, and an adoption score built from real usage (downloads, stars, installs). Each finding says where it is and how to fix it. Packages and MCP servers can also be run in a locked-down sandbox to see what they actually do: network calls, file writes, and whether planted credentials stay put. Every result is signed (Ed25519) and can be recomputed offline against our public keys, so you can check it without trusting us. Read-only. No account. Ask it to check a tool before you install or connect it.
Tools
check_interaction_safety
ClaudeCheck whether it is safe to interact with another agent by trust threshold. Thresholds: delegate 0.6, trade 0.5, collaborate 0.4, follow 0.1. Returns is_safe, risk_level, the score, and a recommendation. Read-only, no auth.
get_trust_badge
ClaudeGet an embeddable AgentAvow trust badge (SVG) for an entity, with ready-to-paste Markdown and HTML. The badge auto-updates as the score changes. Read-only, no auth.
lookup_identity
ClaudeLook up an AgentAvow entity by W3C DID or display name. Returns the entity's id, name, type, trust score and tier. Read-only, no auth. Use to resolve an identity before checking its trust.
scan_mcp_server
ClaudeScan a live MCP server's tool definitions for tool-poisoning, prompt-injection, invisible-unicode, and manifest-execution risks before your agent connects to it. Returns a 0-100 trust score, a safe / needs-review verdict, findings, and a signed attestation. Read-only; calls the agentavow.com public API.
scan_package
ClaudeScan a published package (npm, PyPI, crates, Docker, or Hugging Face) with AgentAvow. Returns a 0-100 trust score, a safe / needs-review verdict, findings with remediation, and a signed attestation. Also reports repo-vs-artifact drift (files shipped that aren't in the source). Read-only; calls agentavow.com.
scan_repo
ClaudeScan a public GitHub repository with AgentAvow and return whether it is safe for an agent to connect to: a 0-100 trust score, a plain safe / needs-review verdict, the findings behind it (with where and how to fix), and a signed, offline-verifiable attestation. Read-only, no account. Calls the AgentAvow public API at agentavow.com.
verify_trust
ClaudeVerify an AgentAvow entity's trust score. Returns trust_score (0-1), trust_score_pct (0-100), trust_tier, and whether it meets a minimum threshold. Read-only, no auth. Use before interacting with an unknown agent.
