Return summary statistics for this callback.

Get pipe statistics.

Get storage statistics.

Get storage statistics.

Get storage statistics.

Return Agent Shield validation statistics. Returns: Dict with total validations, pass/fail counts, and per-stage brea

Get exporter statistics.

Get statistics about recorded data.

Get statistics about recorded data.

Get statistics about recorded data.

Get conversation summary if available.

Get all registered supervisors

Get supported security levels

Get supported routing strategies

Get supported verification types

Get the current status of a task on a peer agent.

Retrieve an existing thread by ID. Args: thread_id: The thread to retrieve. Returns: The OpenAI thread objec

Get metrics for a specific time window. Args: tool_name: Name of the tool. window: Time window to aggregate over.

Acquire an access token for the given scope.

Return cumulative token usage statistics. Returns: A dict with ``prompt_tokens``, ``completion_tokens``, ``total_toke

Return cumulative token usage statistics. Returns: A dict with ``prompt_tokens``, ``completion_tokens``, ``total_toke

Return cumulative token usage statistics. Returns: A dict with ``prompt_tokens``, ``completion_tokens``, ``total_toke

Retrieve a tool specification from the global registry. Args: name: The unique tool identifier. version: Version constraint (e.g.

Get a registered tool by name. Args: name: Tool name Returns: The tool if found, None otherwise

Get a ToolHandle for executing a tool with all policies applied. The ToolHandle provides a convenient interface for executing tools with auto

Get all tools of a specific type

Retrieve flight recorder logs for a specific trace ID. Useful for debugging and distributed tracing across agent calls.

Retrieve flight recorder logs for a trace ID.

Get the current state of the tiered context. Returns: Current tiered context state with item counts

Get the best available trust bridge. Advanced: Cross-protocol trust translation (A2A/MCP/IATP). Community: Direct passthrough without transla

Return direct trust connections for the given agent.

Get the best available trust decay engine. Advanced: Trust contagion + KL divergence regime detection. Community: Linear decay over time.

Get trust verification footer for A2A messages.

Get the best available trust protocol engine. Advanced: Full IATP with sidecar attestation and capability handshake. Community: Basic policy

Get trust status summary for this agent. Returns: Dictionary with agent DID, current score, threshold, interaction co

Get trust status report for this client.

Retrieve trust score for an agent. Args: agent_did: The agent's DID identifier. Returns: Trust score dict or

Get aggregate trust statistics. Returns: Dictionary of trust statistics

Get all unverified events. Returns: List of unverified events

Get user activity within time window.

Get log of validation decisions

Get all validators for an action type

Return the current version of CaaS.

Get detailed version information about the emk package.

Get the edit history of a file. Shows all edits made to a file, including which agents made changes and when. Useful for understandi

Get the complete Virtual File System state. Returns the entire file system state, useful for debugging or snapshotting the current p

Get history of detected violations

Get the current status of a workflow

Get all working memory contents.

Uses global registry.

Agent that always returns correct answers.

Decorator that enforces governance policy on a PydanticAI tool. Checks tool name allowlist, argument content against blocked patterns, semant

Governance gate executed after every LLM call. Scans the LLM response for blocked patterns and runs ``post_execute`` drift de

Governance gate executed after every tool call. Scans the tool result for blocked patterns and runs drift detection via ``pos

Governance gate executed before every LLM call. Scans the message list for blocked patterns and runs Cedar/OPA ``pre_execute`

Governance gate executed before every tool call. Checks tool allowlist/blocklist, scans arguments for blocked patterns, and r

Governed wrapper around CrewAI agent delegation. Intercepts delegation calls between agents, tracks delegation depth, and enf

Delete a file through the governed kernel.

Governed wrapper around a CrewAI agent's task execution. Intercepts each task execution call, applies pre-execution

Governed wrapper that validates and delegates PydanticAI tool calls.

Governed wrapper around a smolagents tool's forward method. Intercepts the tool invocation, validates the call against the ac

Governed wrapper around an AutoGen function_map entry.

Governed wrapper around a CrewAI tool's run method. Intercepts the tool call, runs pre-execution policy checks,

Governed wrapper around CrewAI memory save operations. Validates content before it is written to crew memory,

Governed wrapper around ``memory.save_context``.

Governed wrapper around a CrewAI task step. Intercepts individual step calls within a task, validates inputs against

Grant a capability to an agent. Authentication: required. The caller must present an ``Ed25519-Timestamp`` header signed by the grantor DID.

Show trust relationships as an ASCII art or Mermaid diagram.

Greet someone.