AI Agents

66 entities and 415 agent-callable operations in agents. Each operation is addressed as {mega_tool}/agents/{entity}/{operation} and carries the same permission check over MCP and REST.

`agent`

Operation	Mega-tool	What it does
`catalog`	`smart_action`	The agent catalog for this organization: the global Agent catalog overlaid with the org's AgentConfig (enabled state + execution stats).
`create`	`entity_crud`	Create a new agent
`delete`	`entity_crud`	Delete a agent
`list`	`entity_crud`	List agents with filtering and pagination
`read`	`entity_crud`	Get detailed agent information
`run`	`agent_command`	Execute an agent immediately: creates its task, starts it, and opens the execution record.
`search`	`entity_crud`	Search agents with fuzzy matching
`status`	`report_query`	Detailed status of one agent: catalog entry, this org's config, queue depth across the org subtree, and recent executions.
`update`	`entity_crud`	Update an existing agent

`agent_config`

Operation	Mega-tool	What it does
`create`	`entity_crud`	Create a new agent config
`delete`	`entity_crud`	Delete a agent config
`enable`	`smart_action`	Enable an agent for this organization, keyed on agent_type.
`list`	`entity_crud`	List agent_configs with filtering and pagination
`read`	`entity_crud`	Get detailed agent config information
`update`	`entity_crud`	Update an existing agent config

`agent_decision`

Operation	Mega-tool	What it does
`create`	`entity_crud`	Create a new agent decision
`delete`	`entity_crud`	Delete a agent decision
`get_transitions`	`workflow_action`	Get available transitions for a agent decision
`list`	`entity_crud`	List agent_decisions with filtering and pagination
`read`	`entity_crud`	Get detailed agent decision information
`search`	`entity_crud`	Search agent_decisions with fuzzy matching
`transition`	`workflow_action`	Transition agent decision to a new status
`update`	`entity_crud`	Update an existing agent decision

`agent_execution`

Operation	Mega-tool	What it does
`cost_attribution_daily`	`report_query`	Daily AgentExecution cost rollup broken down by module, agent_type and execution tier.
`create`	`entity_crud`	Create a new agent execution
`delete`	`entity_crud`	Delete a agent execution
`execution_performance_by_tool`	`report_query`	Latency (avg/min/max/p95) and total cost of AgentExecutions grouped by execution tier.
`history`	`report_query`	Execution history across the org subtree, filterable by agent type, status, tier and outcome, with a summary computed over the whole period.
`list`	`entity_crud`	List agent_executions with filtering and pagination
`read`	`entity_crud`	Get detailed agent execution information
`search`	`entity_crud`	Search agent_executions with fuzzy matching
`update`	`entity_crud`	Update an existing agent execution

`agent_hierarchy`

Operation	Mega-tool	What it does
`create`	`entity_crud`	Create a new agent hierarchy
`delete`	`entity_crud`	Delete a agent hierarchy
`list`	`entity_crud`	List agent_hierarchies with filtering and pagination
`read`	`entity_crud`	Get detailed agent hierarchy information
`search`	`entity_crud`	Search agent_hierarchies with fuzzy matching
`update`	`entity_crud`	Update an existing agent hierarchy

`agent_instance`

Operation	Mega-tool	What it does
`create`	`entity_crud`	Create a new agent instance
`delete`	`entity_crud`	Delete a agent instance
`list`	`entity_crud`	List agent_instances with filtering and pagination
`read`	`entity_crud`	Get detailed agent instance information
`search`	`entity_crud`	Search agent_instances with fuzzy matching
`update`	`entity_crud`	Update an existing agent instance

`agent_memory`

Operation	Mega-tool	What it does
`archive`	`smart_action`	Archive a knowledge memory (kept with its history, excluded from spawn and recall).
`create`	`entity_crud`	Create a new agent memory
`delete`	`entity_crud`	Delete a agent memory
`facets`	`smart_action`	Counts for a persona's Memory tab: knowledge memories (total/live/archived, by type, by index section), legacy slug-less tier rows (total, by tier), dangling…
`history`	`smart_action`	The append-only revision log (newest first) of one knowledge memory, by id or by persona_id + slug.
`list`	`entity_crud`	List agent_memories with filtering and pagination
`read`	`entity_crud`	Get detailed agent memory information
`search`	`entity_crud`	Search agent_memories with fuzzy matching
`update`	`entity_crud`	Update an existing agent memory
`upsert_by_slug`	`smart_action`	Create or update the live knowledge memory <slug> of a persona.

`agent_message`

Operation	Mega-tool	What it does
`create`	`entity_crud`	Create an ordinary, non-routed agent message
`delete`	`entity_crud`	Delete an ordinary, non-routed agent message
`list`	`entity_crud`	List routed, declassified agent messages
`read`	`entity_crud`	Read one routed, declassified agent message
`search`	`entity_crud`	Search routed, declassified agent messages
`update`	`entity_crud`	Update an ordinary, non-routed agent message

`agent_persona`

Operation	Mega-tool	What it does
`clear_rbac_cap`	`smart_action`	Remove a persona's RBAC cap.
`create`	`entity_crud`	Create a new agent persona
`delete`	`entity_crud`	Delete a agent persona
`ensure_workspace`	`smart_action`	Give the persona its Claude-Code workspace (StudioProject, one-to-one).
`explain_rbac_cap`	`smart_action`	Explain whether a persona's cap users would admit a PRISM coordinate (mega_tool/app/entity/operation) or hold a permission codename: per cap user along the…
`export_md`	`report_query`	Export a persona as one Markdown document: front matter (name, role, model, effort, authority), the system prompt, skills, resources and the prompt version.
`import_bundle`	`smart_action`	Apply a bundle file map to a persona.
`instructions_history`	`report_query`	A persona's system-prompt versions (number, sha256, length, when).
`list`	`entity_crud`	List agent_personas with filtering and pagination
`mint_session_key`	`smart_action`	Mint an MCP key bound to a persona for a Claude Code session.
`rbac_cap_status`	`smart_action`	A persona's cap status: off \| enforced \| not_enforced (cap on but at least one active session link is not on a bound key), the caps along its ancestry, and…
`read`	`entity_crud`	Get detailed agent persona information
`render_bundle`	`smart_action`	Render a persona's portable home as {files: {path: text}, assets, manifest, next_cursor}: persona.json, CLAUDE.md, .claude/rules\|agents\|commands\|skills…
`route_preview`	`report_query`	Preview which AI candidate (runtime, model, effort) a turn of this persona would run on for task_type, which candidates are excluded and why (metered_gate…
`search`	`entity_crud`	Search agent_personas with fuzzy matching
`set_rbac_cap`	`smart_action`	Cap a persona, and every persona and session under it, to one user's RBAC rights.
`update`	`entity_crud`	Update an existing agent persona

`agent_reminder`

Operation	Mega-tool	What it does
`ack`	`smart_action`	Mark a reminder handled (acked).
`cancel`	`smart_action`	Cancel a reminder you no longer need (soft: the row stays).
`create`	`entity_crud`	Set a reminder for a persona (the agent's own, durable across restarts and handoffs).
`delete`	`entity_crud`	Cancel a reminder (soft: the row stays with status cancelled).
`fire_now`	`smart_action`	Test path: fire one pending reminder immediately through the same producer as the one-minute beat (event emit, compat wake, status fired) with a test: dedup…
`list`	`entity_crud`	List a persona's reminders (soonest first).
`read`	`entity_crud`	Read one reminder with its body, resolution (how the date was derived) and firing history.
`resolve_when`	`smart_action`	Dry-run a reminder date rule and show how it resolves before committing: returns fire_at (UTC) and the resolution (calendar_key, calendar_effective…
`snooze`	`smart_action`	Push a reminder back: status snoozed with a new fire_at, from scheduled, snoozed or fired (a fired one's inbox event is dismissed; the new firing is a fresh…
`update`	`entity_crud`	Edit a pending (scheduled or snoozed) reminder: title, body, when_text, timezone, fire_at or date_rule (re-resolves the date), human_target, source pointer…

`agent_response`

Operation	Mega-tool	What it does
`create`	`entity_crud`	Create a new agent response
`delete`	`entity_crud`	Delete a agent response
`list`	`entity_crud`	List agent_responses with filtering and pagination
`read`	`entity_crud`	Get detailed agent response information
`search`	`entity_crud`	Search agent_responses with fuzzy matching
`update`	`entity_crud`	Update an existing agent response

`agent_review_item`

Operation	Mega-tool	What it does
`approve`	`smart_action`	Approve a pending agent review item.
`create`	`entity_crud`	Create a new agent review item
`delete`	`entity_crud`	Delete a agent review item
`get_transitions`	`workflow_action`	Get available transitions for a agent review item
`list`	`entity_crud`	List agent_review_items with filtering and pagination
`read`	`entity_crud`	Get detailed agent review item information
`reject`	`smart_action`	Reject a pending agent review item.
`search`	`entity_crud`	Search agent_review_items with fuzzy matching
`transition`	`workflow_action`	Transition agent review item to a new status
`update`	`entity_crud`	Update an existing agent review item

`agent_task`

Operation	Mega-tool	What it does
`create`	`entity_crud`	Create a new agent task
`delete`	`entity_crud`	Delete a agent task
`list`	`entity_crud`	List agent_tasks with filtering and pagination
`pending_workload`	`report_query`	Backlog of pending/claimed/in_progress tasks broken down by assignee, module and priority.
`read`	`entity_crud`	Get detailed agent task information
`search`	`entity_crud`	Search agent_tasks with fuzzy matching
`task_aging`	`report_query`	Distribution of non-terminal AgentTask age in hour-buckets (0-1h, 1-8h, 8-24h, 24h+).
`task_sla_analysis`	`report_query`	SLA analysis: overdue and approaching-deadline AgentTasks grouped by module and status.
`task_summary`	`report_query`	Count of AgentTasks broken down by status, priority and module with optional time bucketing.
`task_type_distribution`	`report_query`	Top-N task types by volume with completion and failure rates.
`update`	`entity_crud`	Update an existing agent task

`ai_task`

Operation	Mega-tool	What it does
`cancel`	`smart_action`	Cancel an AI task run (the run owner, or a holder of agents.manage_ai_task_runs).
`poll`	`smart_action`	Read an AI task receipt (the output is returned only to the run owner).
`registry`	`smart_action`	List registered AI task prompt builders (with args_schema), output schemas and completion subject types.
`run`	`smart_action`	Run a one-shot AI task through a registered prism_callable prompt builder (no raw prompt, no inline schema).

`ai_task_run`

Operation	Mega-tool	What it does
`list`	`entity_crud`	List AI task receipts (own runs; others with agents.view_ai_task_runs, without outputs).
`read`	`entity_crud`	Read one AI task receipt (output only to the run owner).

`app_ai_task`

Operation	Mega-tool	What it does
`cancel`	`smart_action`	Cancel a running app-AI task — marks it CANCELLED, retires the persona instance, and enqueues a terminal close command (the governed cancel path, not a raw…
`create`	`entity_crud`	Delegate a task to an AI-assistant template — creates the persona instance, the delegation record, and the AppAITask row, then dispatches auto-execution (the…
`execute`	`smart_action`	Manually trigger execution of an app-AI task in STARTING state — spawns the terminal agent on an online desktop agent (the same synchronous spawn the REST…
`list`	`entity_crud`	List app_ai_tasks with filtering and pagination
`read`	`entity_crud`	Get detailed app ai task information
`search`	`entity_crud`	Search app_ai_tasks with fuzzy matching

`approval_item`

Operation	Mega-tool	What it does
`add_info`	`smart_action`	Agent answers a 'need more info' request: appends the text to the item's context (dated) and returns the item to pending.
`answer_question`	`smart_action`	Agent answers one question the person asked on the item (a threaded reply).
`create`	`entity_crud`	File something you are waiting on a person to decide.
`decide`	`smart_action`	The person's decision on an item (organization admins only): action approve \| deny (optional note), need_info \| question (note required: what you want to…
`list`	`entity_crud`	List approval items waiting on a person (the orchestrator's 'Waiting on you' queue).
`list_decided_since`	`smart_action`	Poll for the person's responses: every item with a decision event (approved, denied, needs_info, question, reopened) after `since` (ISO 8601), each with those…
`mark_viewed`	`smart_action`	Record that the person opened an item (one 'viewed' event per person per item).
`override`	`smart_action`	An organization admin with agents.override_approval_item decides a routed item without its named approvers (for when one is unavailable): action approve \|…
`read`	`entity_crud`	Read one approval item: full context (markdown), references, links and the whole timestamped history (created, viewed, decisions, questions, answers, info…
`reassign`	`smart_action`	Change the named approvers on one open item: add and/or remove user ids.
`update`	`entity_crud`	Agent edits an item's title, summary, priority or client_waiting flag (true when a client or customer is waiting on the answer, false when it is internal work).
`vote`	`smart_action`	A named approver votes on a routed item: action approve \| deny (optional note) or retract (take your vote back while the item is open).
`withdraw`	`smart_action`	Agent retracts an item (no longer needed).

`approval_policy`

Operation	Mega-tool	What it does
`create`	`entity_crud`	Create an approval policy.
`delete`	`entity_crud`	Switch a policy off (soft: is_active=false).
`list`	`entity_crud`	List the approval policies of an organization (active ones by default): who approves which chief orchestrator's items, per category, and the any/all/quorum…
`read`	`entity_crud`	Read one approval policy with its approvers.
`resolve`	`smart_action`	Dry run of approval routing: for a persona (chief orchestrator) and category, which rule matches and who would be asked to approve, in which mode.
`update`	`entity_crud`	Change a policy's mode, quorum, approvers (the whole list is replaced), notes or is_active.

`chat_conversation`

Operation	Mega-tool	What it does
`create`	`entity_crud`	Create a new chat conversation
`delete`	`entity_crud`	Delete a chat conversation
`list`	`entity_crud`	List chat_conversations with filtering and pagination
`read`	`entity_crud`	Get detailed chat conversation information
`search`	`entity_crud`	Search chat_conversations with fuzzy matching
`update`	`entity_crud`	Update an existing chat conversation

`chat_message`

Operation	Mega-tool	What it does
`create`	`entity_crud`	Create a new chat message
`delete`	`entity_crud`	Delete a chat message
`list`	`entity_crud`	List chat_messages with filtering and pagination
`read`	`entity_crud`	Get detailed chat message information
`search`	`entity_crud`	Search chat_messages with fuzzy matching
`update`	`entity_crud`	Update an existing chat message

`content_filter_rule`

Operation	Mega-tool	What it does
`list`	`entity_crud`	List content_filter_rules with filtering and pagination
`read`	`entity_crud`	Get detailed content filter rule information
`search`	`entity_crud`	Search content_filter_rules with fuzzy matching

`desktop_agent`

Operation	Mega-tool	What it does
`create`	`entity_crud`	Create a new desktop agent
`delete`	`entity_crud`	Delete a desktop agent
`list`	`entity_crud`	List desktop_agents with filtering and pagination
`read`	`entity_crud`	Get detailed desktop agent information
`search`	`entity_crud`	Search desktop_agents with fuzzy matching
`update`	`entity_crud`	Update an existing desktop agent

`desktop_tool_execution`

Operation	Mega-tool	What it does
`create`	`entity_crud`	Create a new desktop tool execution
`delete`	`entity_crud`	Delete a desktop tool execution
`desktop_tool_adoption`	`report_query`	Daily DesktopToolExecution adoption: success/failure counts and error rate per tool_name.
`desktop_tool_latency_health`	`report_query`	DesktopToolExecution latency health: avg/p95/p99 execution time per tool_name.
`list`	`entity_crud`	List desktop_tool_executions with filtering and pagination
`read`	`entity_crud`	Get detailed desktop tool execution information
`search`	`entity_crud`	Search desktop_tool_executions with fuzzy matching
`update`	`entity_crud`	Update an existing desktop tool execution

`evidence_pack`

Operation	Mega-tool	What it does
`download`	`smart_action`	Resolve a completed evidence pack to its data-minimized manifest plus a download_url reference to the RBAC-gated REST endpoint (PRISM returns JSON; the ZIP…
`kickoff`	`smart_action`	Start an AI-assurance evidence-pack export for a date window (async).
`status`	`smart_action`	Poll an evidence-pack export job: pending/in_progress → progress percent; completed → thin manifest projection; failed/expired reported as data (not a handler…

`governance_audit_entry`

Operation	Mega-tool	What it does
`list`	`entity_crud`	List governance_audit_entries with filtering and pagination
`read`	`entity_crud`	Get detailed governance audit entry information
`search`	`entity_crud`	Search governance_audit_entries with fuzzy matching

`governed_agent`

Operation	Mega-tool	What it does
`list`	`entity_crud`	List governed_agents with filtering and pagination
`read`	`entity_crud`	Get detailed governed agent information
`search`	`entity_crud`	Search governed_agents with fuzzy matching

`human_oversight_request`

Operation	Mega-tool	What it does
`list`	`entity_crud`	List human_oversight_requests with filtering and pagination
`read`	`entity_crud`	Get detailed human oversight request information
`search`	`entity_crud`	Search human_oversight_requests with fuzzy matching

`inference_rate_card`

Operation	Mega-tool	What it does
`list`	`entity_crud`	List governed inference pricing cards visible to this organization (global default + org overrides)
`read`	`entity_crud`	Read one inference pricing card with its per-model rate lines

`instance_context_message`

Operation	Mega-tool	What it does
`create`	`entity_crud`	Create a new instance context message
`delete`	`entity_crud`	Delete a instance context message
`list`	`entity_crud`	List instance_context_messages with filtering and pagination
`read`	`entity_crud`	Get detailed instance context message information
`search`	`entity_crud`	Search instance_context_messages with fuzzy matching
`update`	`entity_crud`	Update an existing instance context message

`library`

Operation	Mega-tool	What it does
`favorite`	`smart_action`	Toggle a favorite on a library asset.
`grants`	`smart_action`	Read-only: list grants on an asset.
`install_pack`	`smart_action`	Install a library pack's workflows and skills into the org.
`pending_updates`	`smart_action`	Installed library items whose source_version is behind the repo.
`search`	`smart_action`	Search workflows, skills and prompt templates.

`memory_entry`

Operation	Mega-tool	What it does
`create`	`entity_crud`	Propose a new memory entry (every PRISM create is a proposal, section 6).
`delete`	`entity_crud`	Delete a memory entry the caller owns or has scope rights over.
`list`	`entity_crud`	List memory entries visible to the caller (own user-scope + org-scope).
`read`	`entity_crud`	Get a memory entry.
`reject`	`smart_action`	Reject/withdraw a memory proposal.
`update`	`entity_crud`	Propose an update.

`org_agents`

Operation	Mega-tool	What it does
`overview`	`report_query`	Read-only view of an organization's AI agents: persona roster (name, role, status, authority level, skills), recent runs (mission, personas, start/end…

`organization_config`

Operation	Mega-tool	What it does
`create`	`entity_crud`	Create a new organization config
`delete`	`entity_crud`	Delete a organization config
`get_transitions`	`workflow_action`	Get available transitions for a organization config
`list`	`entity_crud`	List organization_configs with filtering and pagination
`read`	`entity_crud`	Get detailed organization config information
`search`	`entity_crud`	Search organization_configs with fuzzy matching
`transition`	`workflow_action`	Transition organization config to a new status
`update`	`entity_crud`	Update an existing organization config

`organization_spawn`

Operation	Mega-tool	What it does
`create`	`entity_crud`	Create a new organization spawn
`delete`	`entity_crud`	Delete a organization spawn
`list`	`entity_crud`	List organization_spawns with filtering and pagination
`read`	`entity_crud`	Get detailed organization spawn information
`search`	`entity_crud`	Search organization_spawns with fuzzy matching
`update`	`entity_crud`	Update an existing organization spawn

`organization_template`

Operation	Mega-tool	What it does
`create`	`entity_crud`	Create a new organization template
`delete`	`entity_crud`	Delete a organization template
`list`	`entity_crud`	List organization_templates with filtering and pagination
`read`	`entity_crud`	Get detailed organization template information
`search`	`entity_crud`	Search organization_templates with fuzzy matching
`update`	`entity_crud`	Update an existing organization template

`persona_channel`

Operation	Mega-tool	What it does
`create`	`entity_crud`	Create a persona channel.
`delete`	`entity_crud`	Soft-delete a persona channel (it stops being synced to the wake registry).
`link_session`	`smart_action`	ONE call links a Claude Code session to a mail alias: resolves the persona that owns the active mail_wake_alias channel at `address` (or takes persona_id and…
`list`	`entity_crud`	List a persona's channels (email identity, mail-wake alias, signature, voice profile, ...).
`read`	`entity_crud`	Read one persona channel.
`subscribe`	`smart_action`	Ensure (create or update) the event subscription this wake channel owns.
`sync_wake_registry`	`smart_action`	Sync the mail-wake registry file from the DB: aliases[a].sessions = (sessions - previously managed) + the wake-enabled active sessions of personas whose…
`test`	`smart_action`	Spine probe, NOT an inbox delivery test: emits events.test.emitted (the only topic a test may emit) in the channel's source org and returns the platform event…
`unlink_session`	`smart_action`	Inverse of link_session: sets wake_enabled=false on the session's link(s).
`unsubscribe`	`smart_action`	Deactivate the event subscription this wake channel owns (the channel itself is kept).
`update`	`entity_crud`	Update a persona channel (the persona cannot be changed).

`persona_event`

Operation	Mega-tool	What it does
`ack`	`smart_action`	Acknowledge inbox event(s) as handled (forward-only; an acked event is never re-delivered).
`dismiss`	`smart_action`	Dismiss inbox event(s) without handling them (terminal).
`fetch_record`	`smart_action`	Header view of the record an inbox event points at (mail: from, to, subject, received_at; task: title, status, project, link; message: author, created_at…
`inbox_status`	`smart_action`	Is this persona's inbox live?
`list`	`entity_crud`	List a persona's inbox events (oldest first, uuid7 cursor).
`listen_targets`	`smart_action`	The personas (reachable by the caller) with an ACTIVE, wake-enabled session link for this session, with their wake channels.
`post_message`	`smart_action`	Post a message into a persona's internal thread (stored in the persona's org) and emit agents.persona.message_posted so a live listener wakes.
`read`	`entity_crud`	Read one persona inbox event (header fields only; the record itself comes from fetch_record).
`reply`	`smart_action`	Reply to an inbox event, internally only.
`seen`	`smart_action`	Mark inbox event(s) as seen (forward-only; stays in the unacked backlog).
`stream_token`	`smart_action`	Mint a 12 h stream token for ws/persona-events/<token>/ and upsert the PersonaSessionLink of this session (role default orchestrator).

`persona_message`

Operation	Mega-tool	What it does
`list`	`entity_crud`	List the internal message thread of a persona (oldest first, uuid7 cursor).

`persona_resource`

Operation	Mega-tool	What it does
`create`	`entity_crud`	Create a persona resource pointer.
`delete`	`entity_crud`	Soft-delete a persona resource pointer (the target itself is untouched).
`list`	`entity_crud`	List a persona's resource pointers (PMO board, ledgers, report dirs, scripts, logs, ...).
`read`	`entity_crud`	Read one persona resource pointer.
`update`	`entity_crud`	Update a persona resource pointer (the persona cannot be changed).

`persona_session_link`

Operation	Mega-tool	What it does
`checkpoint`	`smart_action`	Point a session link at its latest checkpoint note.
`create`	`entity_crud`	Link a Claude Code session UUID to a persona.
`delete`	`entity_crud`	Remove a persona session link (successors keep their rows; their predecessor becomes empty).
`lineage`	`smart_action`	The ordered handoff chain of a session (every link of the lineage with its placement history) and the current head - 'where has this agent lived'.
`list`	`entity_crud`	List the Claude Code sessions linked to personas (handoff chain, wake subscription, checkpoint pointer).
`read`	`entity_crud`	Read one persona session link.
`reconcile`	`smart_action`	Attach persona session links to the StudioSession whose claude_session_uuid matches (sets studio, seat, studio_session, runtime and, for a confined seat…
`register`	`smart_action`	Register a Claude Code session as a persona's session (idempotent on persona + session_uuid).
`register_handoff`	`smart_action`	Record a context handoff: creates the successor session link (role handoff_successor, same lineage, placement and wake subscription carried over) and ends the…
`relocate`	`smart_action`	Record the INTENT to move a persona session to a studio seat (same session_uuid, same lineage; placement row 'relocated').
`update`	`entity_crud`	Update a persona session link (status, wake subscription, checkpoint pointer, chain).

`platform_config`

Operation	Mega-tool	What it does
`create`	`entity_crud`	Create a platform config.
`delete`	`entity_crud`	Delete a platform config (hard delete — the model carries no soft-delete, mirroring REST).
`effective`	`smart_action`	Get the effective (merged system→org→user) configuration for the caller.
`list`	`entity_crud`	List platform configs visible to the org (own + system scope).
`read`	`entity_crud`	Get a platform config (own org or system scope).
`update`	`entity_crud`	Update a platform config (partial — only provided fields change).

`platform_hook`

Operation	Mega-tool	What it does
`create`	`entity_crud`	Create an event-driven automation hook.
`delete`	`entity_crud`	Soft-delete a platform hook.
`list`	`entity_crud`	List event-driven automation hooks for the org.
`read`	`entity_crud`	Get a platform hook with its action and condition.
`test`	`smart_action`	Simulate what a hook would do with sample event data (no side effects).
`update`	`entity_crud`	Update a platform hook (partial — only provided fields change).

`platform_rule`

Operation	Mega-tool	What it does
`activate`	`smart_action`	Activate a custom rule as the org's primary.
`create`	`entity_crud`	Create a custom instruction rule (source is forced to 'custom'; slug auto-generated).
`delete`	`entity_crud`	Soft-delete a custom rule (system rules cannot be deleted).
`fork`	`smart_action`	Fork a system rule into a custom primary for the org.
`list`	`entity_crud`	List composable instruction rules visible to the org (custom + system).
`primary`	`smart_action`	Get the effective primary rule for the org (custom override, else system).
`read`	`entity_crud`	Get a rule with its instruction content.
`update`	`entity_crud`	Update a custom rule (system rules are immutable — fork them instead).

`platform_ui_config`

Operation	Mega-tool	What it does
`create`	`entity_crud`	Create a new platform ui config
`delete`	`entity_crud`	Delete a platform ui config
`list`	`entity_crud`	List platform_ui_configs with filtering and pagination
`read`	`entity_crud`	Get detailed platform ui config information
`search`	`entity_crud`	Search platform_ui_configs with fuzzy matching
`update`	`entity_crud`	Update an existing platform ui config

`policy_rule`

Operation	Mega-tool	What it does
`create`	`entity_crud`	Create a governance policy rule (idempotent on org/name/agent)
`delete`	`entity_crud`	Delete (soft by default) a governance policy rule
`list`	`entity_crud`	List policy_rules with filtering and pagination
`read`	`entity_crud`	Get detailed policy rule information
`search`	`entity_crud`	Search policy_rules with fuzzy matching
`update`	`entity_crud`	Update an existing governance policy rule

`retrieval_trace`

Operation	Mega-tool	What it does
`list`	`report_query`	List retrieval traces in this organization, newest first (no result rows; read one for those).
`read`	`report_query`	Read one retrieval trace with its ranked result rows (ref, document, section, offsets, passage sha256, scores in basis points).

`review_loop_profile`

Operation	Mega-tool	What it does
`list`	`report_query`	List this organization's review-loop profiles (one per registered profile key), plus the registered profile keys that have no row yet.
`read`	`report_query`	Read one review-loop profile: config, version, the generic + profile JSON schemas and the loop graph (nodes/edges, the exit diamond names the active exit mode).
`update`	`workflow_action`	Update an org's review-loop profile config (merged over the current one, validated against the generic and the profile schema; version + 1).

`review_loop_run`

Operation	Mega-tool	What it does
`cancel`	`workflow_action`	Cancel a review-loop run that has not finished.
`events`	`report_query`	Page the hash-chained ledger of a review-loop run in sequence order.
`list`	`report_query`	List review-loop runs in this organization, newest first.
`read`	`report_query`	Read one review-loop run: status, phase, round, scores, exit reason, the config snapshot and every Required Action with its derived status (open/closed).
`resume`	`workflow_action`	Unstick a review-loop run: re-drive a finished-but-unprocessed AI turn, re-enqueue the phase machine, or retry the phase of a failed run.
`signoff`	`workflow_action`	Signoff review loop run.
`start`	`workflow_action`	Start a run of a registered adversarial review loop (drafter → reviewer → arbiter, then a human sign-off) for this organization.
`verify_chain`	`report_query`	Recompute every link of a run's hash-chained ledger.

`skill`

Operation	Mega-tool	What it does
`activate_preview`	`smart_action`	Which of the caller's runnable skills a piece of text would activate (trigger patterns).
`assign_persona`	`smart_action`	Link a persona to a skill.
`catalog`	`smart_action`	The skill catalog: system skills (all orgs) + this org's custom skills.
`create`	`entity_crud`	Create a composable expertise skill.
`delete`	`entity_crud`	Soft-delete a custom skill (system skills cannot be deleted).
`list`	`entity_crud`	List composable expertise skills for the org.
`publish`	`smart_action`	Publish skill.
`read`	`entity_crud`	Get a skill with its prompt content and linked personas.
`remove_persona`	`smart_action`	Unlink a persona from a skill.
`runnable`	`smart_action`	The org skills this caller's agents receive in this org: the same set a Studio/Terminal v2 conversation gets as .claude/skills/org-<slug>/ (one resolver).
`update`	`entity_crud`	Update a skill (partial — only provided fields change).

`skill_bundle`

Operation	Mega-tool	What it does
`import`	`smart_action`	Import a SKILL.md bundle (.zip of <folder>/SKILL.md skills, or one SKILL.md) from a staged upload.
`status`	`smart_action`	Review status of an imported skill bundle: each skill's version, status, licence and files.

`studio_agent`

Operation	Mega-tool	What it does
`create`	`entity_crud`	Create a Studio agent definition under a project.
`delete`	`entity_crud`	Soft-delete (deactivate) a Studio agent definition.
`list`	`entity_crud`	List Studio agent definitions (optionally scoped to a project).
`read`	`entity_crud`	Get a Studio agent definition.
`update`	`entity_crud`	Update a Studio agent definition (partial).

`studio_command`

Operation	Mega-tool	What it does
`create`	`entity_crud`	Create a Studio slash command under a project.
`delete`	`entity_crud`	Soft-delete (deactivate) a Studio slash command.
`list`	`entity_crud`	List Studio slash commands (optionally scoped to a project).
`read`	`entity_crud`	Get a Studio slash command.
`update`	`entity_crud`	Update a Studio slash command (partial).

`studio_hook`

Operation	Mega-tool	What it does
`create`	`entity_crud`	Create a Studio event hook under a project.
`delete`	`entity_crud`	Soft-delete (deactivate) a Studio event hook.
`list`	`entity_crud`	List Studio event hooks (optionally scoped to a project).
`read`	`entity_crud`	Get a Studio event hook.
`update`	`entity_crud`	Update a Studio event hook (partial).

`studio_memory`

Operation	Mega-tool	What it does
`create`	`entity_crud`	Create a Studio memory file under a project.
`delete`	`entity_crud`	Soft-delete (deactivate) a Studio memory file.
`list`	`entity_crud`	List Studio memory files (optionally scoped to a project).
`read`	`entity_crud`	Get a Studio memory file.
`update`	`entity_crud`	Update a Studio memory file (partial).

`studio_project`

Operation	Mega-tool	What it does
`create`	`entity_crud`	Create a Studio project.
`delete`	`entity_crud`	Soft-delete a Studio project and cascade-deactivate all its children.
`file_tree`	`smart_action`	Return the virtual Claude-Code file tree computed from a Studio project's rows.
`import_folder`	`smart_action`	Import a Claude-Code project folder dump (CLAUDE.md, .claude/settings.json, .claude/agents\|rules\|commands/.md, memory/.md, settings hooks) into a new Studio…
`list`	`entity_crud`	List Studio projects (Claude-Code workspace configs) for the org.
`read`	`entity_crud`	Get a Studio project with its nested agents/rules/commands/memories/hooks.
`render_workspace`	`smart_action`	Render the project's workspace as a {filepath: content} map (what a session materializes).
`update`	`entity_crud`	Update a Studio project (partial — only provided fields change).

`studio_rule`

Operation	Mega-tool	What it does
`create`	`entity_crud`	Create a Studio context rule under a project.
`delete`	`entity_crud`	Soft-delete (deactivate) a Studio context rule.
`list`	`entity_crud`	List Studio context rules (optionally scoped to a project).
`read`	`entity_crud`	Get a Studio context rule.
`update`	`entity_crud`	Update a Studio context rule (partial).

`studio_template`

Operation	Mega-tool	What it does
`clone`	`smart_action`	Clone a Studio template project (and its active agents/rules/commands/memories/hooks) into a new non-template project.
`list`	`entity_crud`	List Studio template projects (cloneable starting points) for the org.
`read`	`entity_crud`	Get a Studio template project with its nested children.

`task`

Operation	Mega-tool	What it does
`claim`	`agent_command`	Atomically claim the highest-priority pending agent task visible to the caller across this organization and its descendants, taking a server-authoritative…
`complete`	`agent_command`	Report an agent task as finished (or failed), writing the AgentExecution audit row and releasing the claim lease.

`task_plan`

Operation	Mega-tool	What it does
`create`	`entity_crud`	Create a new task plan
`delete`	`entity_crud`	Delete a task plan
`list`	`entity_crud`	List task_plans with filtering and pagination
`read`	`entity_crud`	Get detailed task plan information
`search`	`entity_crud`	Search task_plans with fuzzy matching
`update`	`entity_crud`	Update an existing task plan

`team`

Operation	Mega-tool	What it does
`add_member`	`smart_action`	Add (or reactivate) an agent instance as a team member.
`create`	`entity_crud`	Create a multi-agent coordination team.
`delete`	`entity_crud`	Soft-delete a team.
`dissolve`	`smart_action`	Dissolve a team (sets status=dissolved).
`list`	`entity_crud`	List multi-agent coordination teams for the org.
`members`	`smart_action`	List the active members of a team.
`read`	`entity_crud`	Get a team with its members and recent tasks.
`remove_member`	`smart_action`	Remove an agent instance from a team (deactivates the membership).
`update`	`entity_crud`	Update a team (partial — only provided fields change).

`team_task`

Operation	Mega-tool	What it does
`assign`	`smart_action`	Assign a task to an agent instance (omit instance_id to unassign).
`complete`	`smart_action`	Mark a task as completed.
`create`	`entity_crud`	Create a coordination task under a team.
`delete`	`entity_crud`	Soft-delete a coordination task.
`list`	`entity_crud`	List coordination tasks for the org (optionally scoped to a team).
`read`	`entity_crud`	Get a coordination task with its dependencies.
`update`	`entity_crud`	Update a task's status / assignment / dependencies.

`watch_rule`

Operation	Mega-tool	What it does
`create`	`entity_crud`	Create a watch rule: when an allow-listed event happens in your scope, run a workflow in PROPOSAL mode (results land in your review queue; nothing is sent or…
`delete`	`entity_crud`	Delete one of your watch rules (its subscription is deactivated).
`list`	`entity_crud`	List your own watch rules in this organization.
`options`	`entity_crud`	What a watch rule can be built from: allow-listed topics (with the identifier bindings and scope kinds each allows), your watchable mailboxes, workflows you…
`read`	`entity_crud`	Get one of your watch rules.
`scope_options`	`entity_crud`	Search the folders (kind=folder), matters (kind=matter) or deals with coverage tests (kind=deal) you can use as a rule's scope.
`update`	`entity_crud`	Update one of your watch rules (name, scope_filter, workflow_inputs, daily_limit, enabled).

`workflow`

Operation	Mega-tool	What it does
`cancel`	`workflow_action`	Cancel a workflow run (the run owner, or a holder of full access on the workflow).
`create`	`entity_crud`	Create a new workflow
`delete`	`entity_crud`	Delete a workflow
`export`	`workflow_action`	Export a workflow as a4.workflow-export/1 JSON (definition, input schema, skill slugs and prompt text; never document ids, grants, runs or matter ids).
`generate_from_description`	`workflow_action`	Describe-to-build: turn a plain-language description into a draft workflow definition built only from registered step types and validated.
`import`	`workflow_action`	Import an a4.workflow-export/1 payload as YOUR private draft.
`list`	`entity_crud`	List workflows with filtering and pagination
`read`	`entity_crud`	Get detailed workflow information
`run`	`workflow_action`	Start a run of a published workflow.
`search`	`entity_crud`	Search workflows with fuzzy matching
`update`	`entity_crud`	Update an existing workflow
`validate`	`workflow_action`	Validate a workflow definition (a4.workflow/1): DAG shape, registered step types, step configs, one entry step, an output step, resolvable {{variables}}.

`workflow_execution`

Operation	Mega-tool	What it does
`create`	`entity_crud`	Create a new workflow execution
`delete`	`entity_crud`	Delete a workflow execution
`list`	`entity_crud`	List workflow_executions with filtering and pagination
`read`	`entity_crud`	Get detailed workflow execution information
`search`	`entity_crud`	Search workflow_executions with fuzzy matching
`update`	`entity_crud`	Update an existing workflow execution

`workflow_step_type`

Operation	Mega-tool	What it does
`list`	`smart_action`	List the registered workflow step types (key, label, palette group, config JSON Schema, output type, side effects, confinable) that a4.workflow/1 definitions…

`workflow_trigger`

Operation	Mega-tool	What it does
`create`	`entity_crud`	Create a new workflow trigger
`delete`	`entity_crud`	Delete a workflow trigger
`list`	`entity_crud`	List workflow_triggers with filtering and pagination
`read`	`entity_crud`	Get detailed workflow trigger information
`search`	`entity_crud`	Search workflow_triggers with fuzzy matching
`update`	`entity_crud`	Update an existing workflow trigger

Parameter schemas and permission codenames are in the full reference — sign in — or call prism_introspect(dimension="capability_search", query="…") with your key.

In the full reference this application appears under agent_command, entity_crud, report_query, smart_action, workflow_action.