@microsoft/agentos-mcp-server

Read a file by path. Requires minimum trust score of 300. Args: path: File path to read. agent_did: The calling agent's DID (e.g. "did:mesh:abc123"). trust_score: The agent's current trust score (0-1000).

Write content to a file. Requires trust >= 600 and fs_write capability. Args: path: Destination file path. content: Content to write. agent_did: The calling agent's DID. trust_score: The agent's current trust score (0-1000). capabilities: List of agent capabilities (must include "fs_write").

Execute a read-only SQL query. Requires trust >= 800, db_query capability, rate-limited to 10/min. Args: sql: The SQL query to execute (read-only). agent_did: The calling agent's DID. trust_score: The agent's current trust score (0-1000). capabilities: List of agent capabilities (must include "db_query").

Read a file from the filesystem (with governance).

Query a database (with governance).

Call an external API (with rate limiting).

Check if an agent is trusted. Returns the agent's overall trust score, trust level, and all five trust dimensions (competence, integrity, availability, predictability, transparency). Args: agent_did: The DID of the agent to check (e.g. "did:mesh:abc123").

Get a detailed trust score with all 5 dimensions. Dimensions: competence, integrity, availability, predictability, transparency. Each dimension is scored 0-1000. Args: agent_did: The DID of the agent to query.

Initiate a trust handshake with a peer agent. Creates a cryptographic challenge, records the handshake, and returns a signed token the peer can verify. Args: peer_did: The DID of the peer agent to handshake with. capabilities: List of capability strings requested for this session.

Verify that a scope chain from *delegator_did* to *agent_did* is valid. Checks that the agent's DID is known, the delegator's trust score is sufficient, and the requested capability is plausible. Args: agent_did: The DID of the agent claiming delegated authority. delegator_did: The DID of the delegator (parent agent). capability: The capability being delegated (e.g. "read:data").

Record an interaction outcome to update trust scores. Valid outcomes: ``success``, ``failure``, ``timeout``, ``partial``. Positive outcomes raise trust; negative outcomes lower it. Args: peer_did: The DID of the peer agent involved. outcome: Interaction result — one of success, failure, timeout, partial. details: Free-text description of the interaction.

Get this agent's identity info. Returns the server's DID, public key, name, and capabilities.