by imran-siddique · v0.0.1
Cryptographic trust verification for multi-agent workflows using CMVK (Ed25519) identities
This community listing does not yet include every recommended support, privacy, pricing, and permission disclosure. Review the available package permissions before installing.
Available inside your emploidai workspace after installation.
Available inside your emploidai workspace after installation.
Cryptographic trust verification for multi-agent Dify workflows using CMVK (Ed25519) identities.
cd dify-plugin
zip -r agentmesh-plugin.zip .
When configuring the plugin in Dify, you can set:
| Parameter | Description | Default |
|---|---|---|
| Minimum Trust Score | Score (0-1) required for verification | 0.5 |
| Cache TTL | Seconds to cache verification results | 900 |
| Identity Name | Name for this agent's identity | dify-agent |
| Capabilities | Comma-separated list of capabilities | (empty) |
Verify another agent's identity before trusting their data or delegating tasks.
Parameters:
peer_did (required): The DID of the peer agent (e.g., did:cmvk:abc123)peer_public_key (required): Base64-encoded Ed25519 public keyrequired_capabilities: Capabilities the peer must havepeer_capabilities: Capabilities the peer claims to haveCheck if the current agent has permission to execute a workflow step.
Parameters:
workflow_id (required): Workflow identifierstep_id (required): Step identifier within workflowstep_type (required): Type of step (llm, tool, code, etc.)required_capability: Specific capability to checkGet this agent's cryptographic identity to share with other agents.
Parameters:
include_capabilities: Whether to include capabilities listRecord successful or failed interactions to update trust scores.
Parameters:
peer_did (required): DID of the peer agentsuccess (required): Whether the interaction succeededseverity: For failures, how much to decrease trust (0-1)1. Agent A calls "Get Agent Identity" to get its DID and public key
2. Agent A shares identity with Agent B
3. Agent B calls "Verify Peer Agent" with A's DID and public key
4. If verified, Agent B proceeds with collaboration
5. After interaction, Agent B calls "Record Interaction" to update trust score
MIT License - See AgentMesh for details.