Return whether OTel has been bootstrapped.

Check if disclosure has been confirmed for a session.

Return ``True`` when the handle has passed its expiry time.

Return whether these claims are outside their freshness window.

Check if the shadow evaluation confidence meets the threshold. Args: min_confidence: Minimum required confidence score (0.0–1.0).

Check if agent is in stopped state.

Check if this policy is stricter than another.

Check if the user context is still valid.

Get keys matching pattern.

Kill all running agents. Returns count killed.

Extract safety corrections from failure critique. This is called by the Observer when it detects a failure.

Remove an agent from this session.

levenshteinDistance returns the edit distance between two strings.

* List all agents.

* List all policy templates.

Return a snapshot of all tracked agent handles.

List all documents in the store. Returns: List of all documents

List all registered runbooks.

Get list of allowed tools, or None if all tools allowed.

Return all registered bundles.

List directory contents.

List all registered plugins (latest version of each). Args: type_filter: Optional filter by plugin type. Returns:

List all provider slots and their current implementations. Returns a dict of {slot_name: "advanced" | "community"}.

List available resources for an agent.

Return all currently revoked agents (excludes expired entries).

List all managed rollouts.

List all registered scanner names.

List recent traces

List stored traces with optional filtering.

Simulate LLM quality degradation (incoherent or low-quality responses).

Load policy rules from a YAML file (legacy flat format).

Parse and load a JSON string as a Policy document.

* Load policies from configuration

* Load policies from configuration

Load agent state from checkpoint.

Load system instructions from JSON file.

Parse and register a policy from a JSON string. Args: json_content: Raw JSON policy definition. Returns: The

Load multiple policies from dictionary representations. Returns the number of policies loaded.

Load a governance policy from a dictionary.

Create an evaluator loaded with one specific policy file.

Load SQL policy configuration from a YAML file. Args: path: Path to a YAML file with ``sql_policy`` section. Returns: SQLPol

Log a burn rate alert firing. Args: slo_name: Name of the SLO alert_name: Name of the alert burn_rate: Cu

Log cost trajectory as a W&B run. Args: agent_id: Agent identifier. costs: List of cost values. Returns:

Log a chaos fault injection. Args: experiment_name: Name of the chaos experiment fault_type: Type of fault injected

Log an incident resolution. Args: incident_id: Unique incident identifier duration_seconds: Time to resolution in sec

Log a trust score change.

Push value to head of list. Returns new list length.

Get list range [start, stop].

List directory contents.

Entry point for ``python -m mcp_trust_server`` and the console script.

Mark an agent as orphaned.

Mark an Undo_API as unhealthy (failed health check).

Return whether this evidence matches the expected binding inputs.

Create a metrics endpoint handler. Usage with FastAPI: from fastapi import FastAPI, Response app = FastAPI()

* Generate mock CMVK response for local development/demo

* Generate mock CMVK response for local development/demo

Simulate a network partition isolating agents from each other.

Create an evaluator that penalizes harmful content.