Record the expected content hash for a tool.

Register a kernel implementation. Args: kernel: The kernel instance to register plugin_id: Optional

Register parties involved in a matter.

Register a policy provider

Register automatic response actions for a signal type.

Register 'sign' subcommand and its sub-subcommands on the main parser.

Register a supervisor agent

Register a trust-gated tool. Args: tool: The tool to register

Register a tool function for automatic execution.

Register a validator. Args: validator: The validator instance plugin_id: Optional ID (defaults to va

Reject a pending agent request.

Mark escalation as rejected.

Release a lock.

No-op in Public Preview.

Release a sponsorship bond.

Replace the loaded trust policies with an exclusive write lock. Acquires the write side of the RW lock so that no evaluations can pro

Mark a violation as remediated. Args: violation_id: ID of the violation to remediate. notes: Free-text description of

Render *report* as a Markdown document. The output is plain Markdown 1.0 — no Markdown extensions, no HTML, so it parses cleanly with every r

Render a polymorphic response as JSON component spec. Args: response: The polymorphic response from the agent

Replay fixtures against policies and return a report. Args: policy_path: Directory or single YAML file containing policies. fixtu

Replay all registered agent traces and check cross-agent consistency. Iterates agents in execution order, replays each with a local R

Callback function called for each replayed event

Report task outcome to update reputation.

Report a task outcome for reputation tracking.

Request approval for a policy-gated action. Implementations may block (waiting for human input), call an external service, or auto-de

Send a TrustRequest. Not yet wired to a real gRPC stub. Raises NotImplementedError. The previous body synthesized a success-shaped re

Require access or raise AccessDeniedError. Args: principal: The principal requesting access. tool_name: Name of the t

Decorator to require trust verification for an endpoint. Args: min_score: Minimum trust score required requir

Reset all counters (for testing).

Reset all internal state (for testing).

Reset call counter for a new session.

Reset call counter (call between requests).

Manually reset the circuit breaker to closed state

Reset the rate limiter state.

Clear state for a single agent and drop its tracked entries. Args: agent_id: Agent identifier whose bucket should be reset.

Clear escalation history for a conversation.

Clear state for a conversation.

Reset counters and start time (for new execution runs).

Clear validation history.

Clear all recorded timestamps for *agent_id* (useful in tests).

Clear all records.

Clear all records.

* Reset a specific agent back to healthy status.

Reset the call counter for *agent_id*. Args: agent_id: Agent identifier whose budget should be reset.

Clear state for every agent. This removes all retained timestamps from every tracked bucket and drops the per-agent lock and tracked-

Reset every circuit breaker.

Zero the four host-tracked budget counters.

Reset per-agent call counters.

Manually reset the restart count for an agent

Reset cache statistics.

Reset statistics.

* Resolve a risk ID to its canonical ASI identifier. * Returns the ID unchanged if it already starts with "ASI", or maps it * via {@link LEGACY_AT_T

Load a policy bundle from a local YAML file. Args: path: Path to the YAML policy file. Returns: The loaded P

Load a policy bundle from Azure Key Vault. The secret value should be a JSON-serialized policy bundle. Args: vault_url:

Load a policy bundle from an HTTP URL. Args: url: The URL to fetch the bundle from (JSON or YAML). Returns:

Perform X3DH as the responder. Args: peer_identity_key: Initiator's X25519 identity public key. ephemeral_public_key:

Respond to simulation result MUTE PATTERN: Returns None if no action needed

Restore a checkpoint — destroy current container, recreate from image.