provenance-mcpSAFE
The three.ws agent action-provenance log over MCP — append-only, ERC-191-signed, on-chain-verifiable. Any AI agent can record what it did and audit another agent's provenance trail. Cursor-paginated reads + a signed append to an immutable ledger.
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
@three-ws/provenance-mcp
The agent action-provenance log over MCP — append-only, ERC-191-signed, on-chain-verifiable. Record what an agent did, and audit another agent's trail, from any AI assistant.
A Model Context Protocol server for the three.ws agent trust layer. Every action an agent takes is written to an append-only ledger — optionally signed with the agent's EVM wallet (ERC-191), so its authorship is provable, not just claimed. This server lets one agent record its own actions and lets any agent read and cryptographically verify another agent's provenance trail.
Why it matters: a forked, purchased, or delegated agent is only trustworthy if its history is provable. provenance-mcp turns "the agent says it did X" into "the agent signed that it did X, and anyone can ecrecover the signer to confirm." The ledger is immutable — records are never edited or deleted.
Install
npm install @three-ws/provenance-mcp
Or run with npx (no install):
npx @three-ws/provenance-mcp
Quick start
**Claude
e08d6e57e344OBSERVED · 2026-10-10Connect
Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control. Replace the environment placeholders with a token scoped to the least it needs.
claude mcp add provenance-mcp --env THREE_WS_TOKEN=${THREE_WS_TOKEN} --env THREE_WS_SIGNER_KEY=${THREE_WS_SIGNER_KEY} -- npx -y @three-ws/[email protected]Trust audit
SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | PASS |
| L1 | Static analysis of the code | PASS |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | PASS |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- none-observed
- Network
- declared (2 observation(s))
- Shell
- none-observed
- Dependencies
- not all pinned
- Secrets in source
- none-found
Findings (3)
.nojekyll
@modelcontextprotocol/sdk, ethers, zod
Gates applied: no_behavioural_pass.
e08d6e57e344full audit observations/trust-audit/mcp-server/nirholas__provenance-mcp.json · Report an issue / request a re-scanAudit history
Every audit this server has had. A grade with a past is a grade somebody is still checking.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-10 | e08d6e57e344 | SAFE | B | 89 | first audit |
Questions
What is the provenance-mcp MCP server?
The three.ws agent action-provenance log over MCP — append-only, ERC-191-signed, on-chain-verifiable. Any AI agent can record what it did and audit another agent's provenance trail. Cursor-paginated reads + a signed append to an immutable ledger.
Is provenance-mcp safe to connect to an agent?
The audit found nothing in the source that contradicts what it says it does, and graded it B (89/100). Grade A is held back for packages that have also passed a sandboxed behavioural run, which is why a clean server reads B.
What credentials does provenance-mcp need?
It reads THREE_WS_SIGNER_KEY and THREE_WS_TOKEN from the environment. Give it a token scoped to the least it needs — an agent that can be talked into calling a tool can be talked into calling it with your credentials.
How does provenance-mcp run?
It speaks stdio, so it runs as a local process your client starts. It is published on npm as @three-ws/provenance-mcp at 0.1.3.
How current is this page?
The grade is for one exact copy of the source (e08d6e57e344), read on 2026-10-10. The repository is watched and re-audited when it changes.