Skip to main content

agentavow-trust

Formerly agentgraph-trust. This distribution is now published as agentavow-trust. The import module is unchanged — import agentgraph_trust still works.

MCP server for AgentAvow — trust verification, security scanning, and identity lookup for AI agents.

Install

uvx agentavow-trust        # zero-install run (recommended)
# or: pip install agentavow-trust

Quick Start

Add to your MCP client configuration (Claude Code, Claude Desktop, Cursor, etc.):

{
  "mcpServers": {
    "agentavow-trust": {
      "command": "uvx",
      "args": ["agentavow-trust"]
    }
  }
}

Then ask your AI assistant:

"Scan the npm package chalk before I install it"
"Scan the repo modelcontextprotocol/servers and tell me if it's safe to connect"
"Scan the MCP server at https://mcp.deepwiki.com/mcp"

This is the local/stdio build of the same service as the remote connector at https://agentavow.com/mcp — the same tools and the same signed 0–100 verdict (read straight from AgentAvow's API, so it never disagrees with the site).

Available Tools

All read-only and anonymous — no account or API key required.

Tool Description
scan_repo Scan a public GitHub repo (owner/name) → Safe to connect / Review before you connect / Do not connect, the reason, a signed 0–100 trust score and the adoption score.
scan_package Scan an npm / PyPI / crates / Docker / Hugging Face package.
scan_mcp_server Scan a live MCP server's tool definitions for poisoning / prompt-injection.
verify_trust Resolve an agent identity and return its current trust score.
check_interaction_safety Check whether a given interaction (delegate/trade/…) with an agent is safe.
lookup_identity Resolve a DID (did:web:… / did:key:…) or search agents by name.
get_trust_badge Get a shields-style trust badge (SVG URL + README markdown) for an agent.
about_agentavow What AgentAvow checks and how to read a verdict.

Every scan result leads with one of three answers — decision = safe (Safe to connect), review (Review before you connect) or do_not_connect (Do not connect) — and decision_reason, the one condition that triggered it ("one high finding: undeclared network call", "nothing found; little code to inspect"). The summary line starts with the phrase (plus " · Certified" when the tool carries the mark). Underneath: the 0–100 trust_score, the tier as detail, certified eligibility, the adoption score (count + unit), the findings behind it, and the binary verdict (safe / needs_review) with its verdict_reason (clean / blocking_findings / thin_coverage / low_signals), unchanged for existing consumers.

Signed attestations

Scan results are cryptographically signed (Ed25519, JWS per RFC 7515) and recomputable offline. Verify signatures against the public JWKS endpoint:

https://agentgraph.co/.well-known/jwks.json

The JWKS host and signing identity intentionally keep the agentgraph.co name so existing attestations stay verifiable; the service and site are AgentAvow at agentavow.com.

Configuration

Env var Default Purpose
AGENTAVOW_URL https://agentavow.com API/site base. AGENTGRAPH_URL is still honored for back-compat.

Metadata

Release files for agentavow-trust 0.6.4

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for agentavow-trust 0.6.4
File Size Uploaded Provenance Actions
agentavow_trust-0.6.4.tar.gz 10.6 kB Without provenance Details for agentavow_trust-0.6.4.tar.gz

Built distribution (wheel)

Table of built distributions (wheels) for agentavow-trust 0.6.4
File Interpreter ABI Platform Actions
agentavow_trust-0.6.4-py3-none-any.whl Python 3 none any Details for agentavow_trust-0.6.4-py3-none-any.whl

Total release size: 21.6 kB

Release files / agentavow_trust-0.6.4.tar.gz

Download URL agentavow_trust-0.6.4.tar.gz
Size 10.6 kB
Tags Source
SHA-256 checksum
How to use checksums
f0a2687e0640656bd3a931d6a22e439c3478c220dc59002aa38361b19e76e278
BLAKE2b-256 checksum
How to use checksums
3f6f77401ceea88c060f7db180c54021f10c4f28f4542b24d95c2755ddab66e2
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.12.13

Release files / agentavow_trust-0.6.4-py3-none-any.whl

Download URL agentavow_trust-0.6.4-py3-none-any.whl
Size 11.0 kB
Tags Python 3
SHA-256 checksum
How to use checksums
0ca57309765c3f08ad7ef441b02025d5f93d85ca8fbee49079388f34aaa84247
BLAKE2b-256 checksum
How to use checksums
d2ef687ddbf0ee64e65214bc71403a8c8324f78d59c922d5380f4ce114c09b40
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.12.13

Security for agentavow-trust 0.6.4

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

No provenance found

None of the files in this release were uploaded with provenance metadata.

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page