Append text to $GITHUB_STEP_SUMMARY.

Persist an entry to file.

Write data to target server stdin.

Access patient data with HIPAA controls.

Delete a message by ID. Returns True if existed.

Mark emotion recognition notice as acknowledged for a session.

Async version of acquire. Args: blocking: If True, wait until a slot is available. timeout: Maximum time to wait if b

Agent attempts to solve the query using current wisdom. Uses prioritization framework to rank context by importance.

Agent attempts to solve the query using its current system instructions.

Map policy action to accent color class.

Log an adapter call at INFO level.

Wire ``agt migrate`` flags onto *parser*.

Add article to knowledge base.

Add a candidate with data consent check.

Add a party that creates a conflict.

Add contract to a matter.

Add a custom policy rule

Add a database table to the graph

Add event to current span.

Add an event to the span.

Add a file path to the graph

Add a job with screening criteria.

Add patient to system.

Add a policy rule to the kernel

Add a principle to the validator.

Register a resource node.

Add a shutdown hook to be called during shutdown

Add a reward signal.

Add a supervisor agent to watch other agents. Args: supervisor_id: ID of the supervisor agent watche

Add a custom tool to ActionType mapping.

Add a conversation turn.

Add a validator to the chain.

Move to the next step in the rollout. Automatically promotes (completes) the rollout if advancing past the last step. Return

Check if command affects protected paths.

ADK after_tool_callback hook for audit logging.

Agent A - Initial thought generator.

Agent B - Tool executor that responds to thoughts.

Agent C - Error handler and logger.

Evaluate and enforce the ``agent_shutdown`` lifecycle point. ``summary`` is the shutdown-summary policy target. A ``deny`` raises :cl

Evaluate and enforce the ``agent_startup`` lifecycle point. ``agent`` is the agent-metadata policy target (e.g. ``{"name": ...}``). A

Get agents with trust scores below a threshold.

Find agent IDs matching a tag key-value pair.

Create an AI-based customer service agent.

Create an AI-based date parser. This is the "New World" - we don't write parseDate() anymore. We let the AI handle it and constrain

Create a policy that allows all access.

Analyze contract content for risky clauses.

* Analyze code for policy violations

Handle analyze command.

Analyze patient data (aggregate, anonymized).

Analyze draft against all lab reports Finds contradictions in: - Efficacy claims - Dosage recommendations

Analyze a manifest and extract ring hints and actions.

Analyze data for privacy concerns and PII. Args: data: Data to analyze Returns:

Analyze query for dangerous patterns.

Analyze execution patterns for risk trends.

Evaluate the current step's analysis criteria against metrics. Args: metrics: Dict mapping metric names to their current values

Analyze a single execution trace and determine if learning is needed. Returns analysis results or None if no learning needed.

Check if this policy applies to a given agent. A policy applies when the agent DID matches ``self.agent``, appears in ``self.agents``

Apply incident/recovery/normal drift to core metrics.

* Simple regex-based rule application.