MCP tool catalog (generated)

MCP tool catalog (generated)

Intended readers: Agent tool developers, orchestrator contributors, and client integration authors.

Machine-derived from contracts/mcp/tool-registry.canonical.yaml.

Total tools: 373

Tool	Product Lane	Tier	HTTP Read Eligible	Description
vox_a2a_ack	ai		No	Acknowledge a message in an agent’s inbox.
vox_a2a_broadcast	ai		No	Broadcast an A2A message to all agents.
vox_a2a_history	ai		No	Query the A2A message audit trail.
vox_a2a_inbox	ai		No	Read unacknowledged messages in an agent’s inbox.
vox_a2a_send	ai		No	Send a targeted A2A message from one agent to another.
vox_a2a_tasks	ai		No	Returns the full DAG of current orchestrator tasks.
vox_add_approval_allowlist_entry	workflow		No	Persist a per-repository “always allow this tool” entry so it auto-approves without parking for HITL approval (T0.3).
vox_agent_assess	ai		No	Evaluates remaining work, returns completion estimate.
vox_agent_continue	ai		No	Triggers auto-continuation for idle agents.
vox_agent_events	ai		No	Streams event history for agents.
vox_agent_gateway_call	interop		No	Invoke an arbitrary agent Gateway method with JSON params.
vox_agent_handoff	ai		No	Passes plan/context from one agent to another.
vox_agent_list_remote	interop		No	List remote agent skills through the native agent runtime adapter.
vox_agent_notify	interop		No	Send a domain-scoped notification over the agent gateway.
vox_agent_status	ai		No	Returns current agent state, activity, mood, queue depth.
vox_agent_subscribe	interop		No	Subscribe the agent gateway session to a domain.
vox_agent_subscriptions	interop		No	List active agent gateway subscriptions for the current session.
vox_agent_unsubscribe	interop		No	Unsubscribe the agent gateway session from a domain.
vox_agy_delegate	ai		No	Delegate a code-generation or refactoring task to Gemini via the agy CLI. Runs agy -p in an isolated git worktree (jail), auto-accepts all prompts, retries on quota/timeout, and auto-logs to the handoff ledger. Returns the diff for human review.
vox_agy_delegate_batch	ai		No	Fan out multiple file-disjoint agy delegation tasks with bounded concurrency (≤8 workers). Each worker gets its own worktree and ledger entry. All results returned for review before integration.
vox_agy_doctor	ai		Yes	Detect whether the Antigravity agy CLI is installed and ready; returns status plus exact install/auth remediation. Call before vox_agy_delegate.
vox_agy_ledger_digest	ai		Yes	Read the handoff ledger and return historical Gemini failure-category frequencies + entry count — the flywheel input injected as explicit avoid-rules into the next launch statement (Stage 1).
vox_agy_pipeline	ai		No	Run one Antigravity delegation end-to-end with deterministic verification: delegate via agy in a worktree jail, capture diff + spend proxy, run the plan-specified gates INSIDE the jail, classify green/partial/failed by effect, write a provisional ledger entry, and return a verdict-ready report for the human merge gate. Use this (not vox_agy_delegate) when you want the result proven, not asserted.
vox_agy_review	ai		No	Record the Claude-side adversarial code review of a delegation as an append-only {id}-review ledger addendum (verdict + §B failure categories + findings + prompt lessons). Closes the review phase before the human merge gate and feeds the flywheel.
vox_answer_question	ai		No	Answer a pending question from another agent.
vox_apply_structured_edit	ai		No	Apply a structured file edit with validation (.vox AST/HIR, .rs shadow cargo check) and bounded LLM repair on failure.
vox_ask_agent	ai		No	Ask another agent a question.
vox_ask_clarification	ai		No	Raise a clarification question to the user and declare gating task IDs.
vox_attention_history	ai		No	Returns the history of attention shifts.
vox_attention_reset	ai		No	Resets the attention budget for the agent pool.
vox_attention_summary	ai		Yes	Returns a summary of agent attention allocation and focus shifts over a time window.
vox_behavior_record	ai		No	Record a user behavior event and receive pattern suggestions.
vox_behavior_summary	ai		No	Analyze recent behavior and summarize detected patterns.
vox_benchmark_list	platform		No	List recent benchmark_event rows from Codex for this repository (requires VoxDb). Also covers build times and eval scores. Trust/sensitivity SSOT — docs/src/architecture/telemetry-trust-ssot.md.
vox_benchmark_record	platform		No	Record a new benchmark observation (e.g. wall-clock time for a build phase). Trust/sensitivity SSOT — docs/src/architecture/telemetry-trust-ssot.md.
vox_broadcast	ai		No	Broadcast a message to all agents on the board.
vox_browser_act	ai		No	LLM proposes a single JSON action (click/fill/goto/wait/noop); host executes it on the page.
vox_browser_back	interop		No	Navigate to the previous history entry for page_id.
vox_browser_click	interop		No	Click an element by CSS selector or xpath:/… target.
vox_browser_click_ref	interop		No	Click an element by ref from the latest accessibility snapshot; use snapshot refs, not CSS selectors. Callers must check data.ok and data.needs_human because sensitive refs return data.ok=false.
vox_browser_click_xy	interop		No	Click at viewport coordinates (x,y) on an existing page_id session.
vox_browser_close	interop		No	Close a browser tab by page_id and shut down the host when no tabs remain.
vox_browser_cookies_export	interop		No	Export cookies to a file under the browser profiles directory. Returns count and path only — cookie values never appear in the MCP payload.
vox_browser_cookies_import	interop		No	Import cookies from a cookies.json file jailed under the browser profiles directory. MCP result has no cookie values.
vox_browser_extract	ai		No	Use configured chat model + visible page text to answer an extraction instruction.
vox_browser_extract_json	ai		No	Like vox_browser_extract but forces a JSON object matching a JSON Schema description string.
vox_browser_fill	interop		No	Focus an input and type text (CSS selector or xpath:/…).
vox_browser_fill_ref	interop		No	Fill an element by ref from the latest accessibility snapshot; use snapshot refs, not CSS selectors. Callers must check data.ok and data.needs_human because sensitive refs return data.ok=false.
vox_browser_forward	interop		No	Navigate to the next history entry for page_id.
vox_browser_goto	interop		No	Navigate an existing tab to a URL.
vox_browser_html	interop		No	Read HTML for the full document (empty target) or a fragment selector.
vox_browser_list_pages	interop		No	List all currently open browser sessions with page_id, URL, and title.
vox_browser_open	interop		No	Open a Chromium tab at URL via CD (chromiumoxide). Returns page_id for follow-up tools.
vox_browser_open_ex	interop		No	Open a Chromium tab with launch mode ephemeral, named, or attach. Named mode requires save_profile or stored cookie consent. Empty VOX_BROWSER_ALLOWED_HOSTS with Named or Attach is an operator risk — set the var in production.
vox_browser_page_info	interop		No	Get URL/title and history navigation booleans for page_id.
vox_browser_press	interop		No	Press a key/chord on the active page.
vox_browser_reload	interop		No	Reload the current page_id.
vox_browser_screencast_frame	interop		No	Capture one CDP screencast frame, persist under browser-frames cache, return path and viewport dimensions.
vox_browser_screenshot	interop		No	Save a full-page PNG to a path (creates parent dirs).
vox_browser_screenshot_viewport	interop		No	Capture viewport-only PNG, persist under browser-frames cache, return path and dimensions.
vox_browser_scroll	interop		No	Scroll by dx/dy viewport deltas.
vox_browser_set_control_lock	interop		No	Set or clear per-page control owner (human or agent) for shared CDP sessions.
vox_browser_set_viewport	interop		No	Set viewport width/height for deterministic coordinate mapping.
vox_browser_snapshot	interop		No	Capture a compact accessibility snapshot with stable element refs for ref-based actions; use these refs instead of CSS selectors.
vox_browser_stop	interop		No	Stop pending navigation/load for page_id.
vox_browser_text	interop		No	Read inner text for a CSS or xpath target.
vox_browser_type	interop		No	Type text into the focused editable element on page_id.
vox_browser_wait_for	interop		No	Wait until a selector appears or timeout (seconds).
vox_budget_status	ai		No	Returns token usage and approximate costs across all agents.
vox_build_crate	app		No	Run cargo build for a crate or the whole workspace.
vox_builtin_registry	app		No	Return detailed signatures and docs for all builtin functions.
vox_cancel_task	ai		No	Cancels an active or queued task.
vox_capability_explain	platform		No	Explain a Vox capability (what it is, its owning crate and layer, CLI/GUI/MCP equivalents and docs link), answered from contracts/capabilities/capability-index.v1.json.
vox_capability_features	platform		No	List Vox capabilities by domain with layer, CLI/GUI/MCP equivalents and docs path, answered from contracts/capabilities/capability-index.v1.json.
vox_capability_model_manifest	platform		No	Return the live Vox capability model manifest JSON (MCP tools, CLI paths, runtime builtin maps) merged from contracts/capability and registries at the workspace root.
vox_change_create	ai		No	Start tracking a new logical change.
vox_change_log	ai		No	Show the history of a change.
vox_chat_history	ai		No	Retrieve the full chat history for the current session.
vox_chat_message	ai		No	Send a chat message to the Vox AI. Resolves @mentions, injects editor context, queries LLM, persists history.
vox_chat_session_append	ai		No	Append one message to a GUI chat session’s display transcript, creating the session on first use.
vox_chat_session_archive	ai		No	Archive a GUI chat session, or restore an archived one with archived=false.
vox_chat_session_create	ai		No	Create a GUI chat session (display transcript) tagged with the server repository.
vox_chat_session_rename	ai		No	Rename a GUI chat session.
vox_check	app		No	CLI operation vox check
vox_check_file_owner	ai		No	Check which agent currently owns a given file path.
vox_check_mood	ai		No	Returns the current gamification mood and status of the agent companion.
vox_check_workspace	app		No	Run cargo check for the entire workspace and return diagnostics.
vox_claim_file	ai		No	Request ownership of a specific file.
vox_code_audit_findings_upsert	platform		No	Record or update TOESTUB anti-pattern findings from external reviews (GitHub/CodeRabbit).
vox_compaction_status	ai		No	Get current context token usage and whether compaction is recommended.
vox_compiler::ast_inspect	app		No	Parse a .vox file and return its AST as a JSON tree. Argument: path (relative to repo root).
vox_complete_task	ai		No	Mark a task as completed, releasing its file locks.
vox_config_get	platform		No	Get the current runtime orchestrator and toolchain configuration (wire aliases: vox_get_config).
vox_config_set	platform		No	Update the orchestrator configuration dynamically (wire alias: vox_set_config).
vox_conflict_diff	ai		No	Show the N-way diff of a conflict.
vox_conflicts	ai		No	List active file conflicts between agents.
vox_context_budget	ai		No	Get the token budget status and summarize recommendation for an agent.
vox_cost_history	ai		No	Get a time-series cost breakdown of operations.
vox_coverage_report	app		No	Get code coverage report for a crate using cargo-llvm-cov.
vox_credentials_status	ai		Yes	List every inference provider with a present Clavis credential plus the agy delegation status — the credential-aware model-selection surface. OpenRouter is one of many.
vox_db_conversation_edge_insert	data		No	Insert conversation_edges between two conversations (V17).
vox_db_conversation_version_append	data		No	Append conversation_versions for a conversation_id (V17).
vox_db_data_flow	data		No	Return the data flow map.
vox_db_explain_query	data		No	Explain a query or mutation in plain English.
vox_db_relationships	data		No	Return the entity-relationship graph for the database.
vox_db_research_metric_linked	data		No	Upsert research_sessions then append research_metrics with matching session_id text (links structured + legacy telemetry).
vox_db_research_session_upsert	data		No	Upsert research_sessions by session_key (V17). Empty repository_id defaults to workspace repository_id.
vox_db_sample_data	data		No	Fetch sample data from a given database table.
vox_db_schema	data		No	Return the complete database schema digest as JSON.
vox_db_suggest_query	data		No	Suggest the correct Vox query expression for an intent.
vox_db_topic_evolution_append	data		No	Append topic_evolution_events for a topic_id (V17).
vox_db_trust_drift	data		No	Compare mean trust observation scores in a recent time window vs the prior equal-length window (optional entity_type/dimension filters).
vox_db_trust_propagate	data		No	Domain-clique affinity propagation over model trust rollups for a base dimension; optional persist writes *_propagated observations.
vox_db_trust_rollups	data		No	List trust_rollups rows with optional entity_type/dimension/domain filters for multidimensional trust inspection.
vox_db_trust_summary	data		No	Aggregate trust_rollups by dimension, domain, entity_type, or combined keys (mean/min/max score, counts, sample mass).
vox_decorator_registry	app		No	Return detailed metadata for all @decorators in the Vox language.
vox_doubt_task	ai		No	Mark a task as suspicious, triggering a rigorous ResolutionAgent audit and potentially unlocking doubt-related achievements.
vox_drain_agent	ai		No	Remove all queued tasks from an agent without retiring it.
vox_emergency_stop	ai		No	Trigger a global emergency stop across the orchestrator, cancelling all active tasks.
vox_exec_time_query	ai		Yes	Query historical execution time for a tool.
vox_exec_time_record	ai		No	Record an execution duration and outcome.
vox_export_grammar_ebnf	app		No	Export the Vox EBNF grammar.
vox_fail_task	ai		No	Mark a task as failed with a reason string.
vox_feedback_list	ai		Yes	List all pending and withheld feedback requests.
vox_file_graph	ai		No	Get a JSON graph of all files and their owning agents (affinity map).
vox_gamify_battle_start	ai		No	Start a bug battle from a synthetic finding (companion name + rule + message).
vox_gamify_battle_submit	ai		No	Submit code for the active bug battle for a companion (success flag).
vox_gamify_collegium_join	ai		No	Join a collegium and emit a Ludus collegium_joined policy event.
vox_gamify_companions	workflow		No	The canonical user’s companions with a server-rendered mood SVG (loopback-only).
vox_gamify_discovery_record	workflow		No	Record a Console discovery exposure (seen or used) for the canonical user (loopback-only).
vox_gamify_due_actions	workflow		No	List discovery actions due for resurfacing for the canonical user (loopback-only).
vox_gamify_leaderboard	workflow		No	Top Ludus players by XP for Axis (loopback-only).
vox_gamify_notification_ack	ai		No	Mark one Ludus notification as read for the canonical user (Codex gamify_notifications), by notification id.
vox_gamify_notifications_ack_all	ai		No	Mark all unread Ludus notifications as read for the canonical user.
vox_gamify_notifications_list	ai		No	List unread Ludus (gamification) notifications for the current canonical user from Codex.
vox_gamify_profile	workflow		No	Read the canonical user’s Ludus profile (regenerating energy) for Axis (loopback-only).
vox_gamify_progress_snapshot	ai		No	Ludus digest: KPI summary, unread notifications, and recent reward-policy snapshots over a rolling day window.
vox_gamify_quest_list	ai		No	List Ludus quests for the canonical user (read-only).
vox_gamify_record_event	workflow		No	Route one Axis GUI event through the Ludus reward policy (loopback-only).
vox_gamify_shop_buy	ai		No	Buy a shop item by 1-based index; optional idempotency_key dedupes duplicate purchases.
vox_gamify_shop_catalog	ai		No	Crystal shop catalog with effective prices for the current gamify mode.
vox_generate_code	ai		No	Generate validated Vox code from a prompt; optional output_path (repo-relative) writes UTF-8 and returns meta.file_outcomes; optional vcs_agent_id captures a post-write snapshot id in meta.file_outcomes.post_write_snapshot_id.
vox_get_active_model	ai		No	Show sticky MCP chat override and resolved ModelSpec (no LLM call).
vox_get_context	ai		No	Retrieve a value from the shared context.
vox_ghost_text	ai		No	LLM fill-in-the-middle completion for single-line editor fragments.
vox_git_blame	app		No	Show line-by-line git blame for a file.
vox_git_diff	app		Yes	Show uncommitted git diff for a file or the whole tree.
vox_git_log	app		No	Show recent git commits (default: last 10).
vox_git_status	app		Yes	Get current git working tree status.
vox_gui_components	ai		Yes	Retrieve the components registered in the primitive component registry.
vox_gui_preference_get	data		No	Read one Axis gui.* preference from the local workspace database (loopback-only).
vox_gui_preference_set	data		No	Write one Axis gui.* preference to the local workspace database (loopback-only).
vox_gui_rules	ai		Yes	List the registered GUI design rules (gui-design-rule/*) so a generator can read the constraint set before emitting.
vox_gui_run_finish	workflow		No	Mark a GUI run completed or failed, joining approval and telemetry totals when a task id is given.
vox_gui_run_start	workflow		No	Record a running GUI-initiated run in the agent_runs journal.
vox_gui_tokens	ai		Yes	Retrieve design tokens in W3C DTCG format.
vox_handoff_context	ai		No	Handoff summarized context from one agent to another.
vox_handoff_lineage	ai		Yes	Returns the historical lineage of agent-to-agent task delegations and context handoffs.
vox_harness_corpus_scan	data		No	Scan the server workspace golden corpus for stale frontmatter and record new harness issues.
vox_harness_fix_propose	ai		No	Ask the LLM facade for a corrected file for a harness issue and store it as a pending fix proposal.
vox_harness_fix_resolve	data		No	Approve (write the proposed content verbatim) or reject a pending harness fix proposal.
vox_harness_issue_decide	data		No	Record a human confirm/dismiss decision for a harness issue.
vox_heartbeat	ai		No	Send an active heartbeat from a Vox agent session.
vox_hopper_cancel	workflow		No	Cancel a hopper item, removing it from the open queue.
vox_hopper_mark_done	workflow		No	Mark a hopper item completed.
vox_hopper_reprioritize	workflow		No	Override a hopper item’s priority (0 background, 1 normal, 2 urgent) with a developer capability.
vox_hopper_submit	workflow		No	Submit an intent to the persistent task hopper as a developer-sourced item.
vox_inline_edit	ai		No	AI inline edit on a file range. Editor sends current text; Rust queries LLM and returns replacement.
vox_journey_canonical_steps	ai		Yes	List ordered canonical developer-journey steps from Codex (SSOT developer_journey_steps).
vox_kb_cite	ai		Yes	Build a citation (heading path, doc link with anchor, kb:// URI, indexed commit) for an indexed KB Library chunk by chunk_hash.
vox_kb_get	ai		Yes	Fetch one indexed KB Library document by docs/src rel_path (frontmatter, chunk list, reconstructed markdown) or one chunk by 64-hex chunk_hash; reads only the index, never the file system.
vox_kb_import_research_session	data		No	Persist a research session’s citations and report into the knowledge base.
vox_kb_ingest_text	data		No	Ingest pasted text into the knowledge base as a document node keyed by content hash.
vox_kb_ingest_url	data		No	Scrape a web page (honouring robots.txt) and ingest it into the knowledge base.
vox_kb_node_delete	data		No	Delete one knowledge-base node and its incident edges (Axis Memory surface).
vox_kb_reindex	ai		No	Rebuild the KB Library index from the repository docs/src tree (takes no path arguments; single-flight); returns the index report.
vox_kb_search	ai		Yes	Search the curated KB Library (Vox docs plus docs/src/library) by BM25 over heading chunks; returns ranked hits with chunk_hash, heading_path, anchor, snippet and kb:// URIs.
vox_kb_status	ai		No	Report KB Library index state, covering corpus root and commit, doc/chunk/card counts, chunker version, lexical backend, embed model, trust tier, per-file errors and index consistency.
vox_knowledge_query	data		No	Query the knowledge graph (VoxDB) for related concepts by keyword.
vox_language_surface	app		No	Return all primary keywords, decorators, types, and builtins in the Vox language.
vox_learn_pattern	ai		No	Record a learned behavioral pattern with confidence score.
vox_lint_crate	app		No	Run cargo clippy for a crate or whole workspace.
vox_list_approval_allowlist	workflow		No	List MCP tools currently allowlisted for auto-approval in a repository (T0.3).
vox_list_context	ai		No	List available context keys by prefix.
vox_list_models	ai		No	List all models in the orchestrator registry (ids, providers, free/paid).
vox_lock_status	ai		No	Returns a list of all current file locks.
vox_map_agent_session	ai		No	Map a Vox agent session ID to an existing orchestrator agent (wire aliases: vox_map_opencode_session, vox_map_vscode_session).
vox_memory_list_keys	ai		No	List all section keys from MEMORY.md.
vox_memory_log	ai		No	Append an entry to today’s daily memory log.
vox_memory_recall	ai		No	Retrieve a fact from long-term memory by key.
vox_memory_recall_db	ai		No	Recall typed memory facts for an agent from VoxDb.
vox_memory_reindex	data		No	Two-way sync between the daemon’s MEMORY.md store and the memory database.
vox_memory_save_db	ai		No	Persist a typed memory fact to VoxDb agent_memory table.
vox_memory_search	ai		No	Hybrid retrieval over memory logs, VoxDb document chunks, and repo inventory (vox memory search; MCP vox_memory_search).
vox_memory_store	ai		No	Persist a key-value fact to long-term memory (MEMORY.md).
vox_mesh_dispatch	workflow		No	Dispatch a .vox script or bundle to the Vox Populi control plane for remote execution on a mesh worker (requires configured control URL).
vox_mesh_nodes	workflow		Yes	List Vox Populi mesh nodes (control plane preferred, local-registry fallback) with status, GPU, trust tier, advertised models.
vox_mesh_queue_stats	workflow		Yes	Vox Populi mesh pending-queue stats (control plane preferred, local queue_depth fallback).
vox_models_preference_set	ai		No	Persist one Models/Settings preference (active_model, routing_priority, selection_policy) for the local GUI user; other keys are refused and values are validated.
vox_my_files	ai		No	Returns all files currently owned by the specified agent.
vox_news_approval_status	data		No	Return distinct approver count and whether dual approval is satisfied for a news_id.
vox_news_approve	data		No	Record a maker-checker approval for a news_id in VoxDb (two distinct approvers required before live syndication).
vox_news_draft_research	data		No	Writes docs/news/drafts/{id}.md from the embedded research template (dry_run in frontmatter).
vox_news_simulate_publish_gate	data		No	Parse news markdown and report what would block live publish (dry_run, approvals, armed) without posting.
vox_news_test_syndicate	data		No	Validates a markdown string against the UnifiedNewsItem parser and executes a pure dry_run without posting anything live.
vox_openclaw_discover	interop		No	Resolve OpenClaw gateway/catalog endpoints from well-known discovery plus local overrides.
vox_openclaw_gateway_call	interop		No	Invoke an arbitrary OpenClaw Gateway WS method with JSON params.
vox_openclaw_health	interop		No	Probe OpenClaw HTTP and WS readiness using the shared discovery configuration.
vox_openclaw_import_skill	interop		No	Import an OpenClaw skill by slug and optionally install it into the local Vox skills registry.
vox_openclaw_list_remote	interop		No	List remote OpenClaw skills through the native OpenClaw runtime adapter.
vox_openclaw_notify	interop		No	Send a domain-scoped notification over the OpenClaw gateway.
vox_openclaw_search_remote	interop		No	Search remote OpenClaw skills by keyword query.
vox_openclaw_subscribe	interop		No	Subscribe the OpenClaw gateway session to a domain.
vox_openclaw_subscriptions	interop		No	List active OpenClaw gateway subscriptions for the current session.
vox_openclaw_unsubscribe	interop		No	Unsubscribe the OpenClaw gateway session from a domain.
vox_oplog	ai		No	Show recent operations with undo support.
vox_oratio_listen	ai		No	Sessionized Oratio transcription with timeout/profile controls. Arg: path plus optional session fields.
vox_oratio_status	ai		No	Oratio / Candle Whisper backend status, decode runtime, and streaming endpoint metadata (JSON; includes stream_ws_url).
vox_oratio_transcribe	ai		No	Thin STT: transcribe one file to text (deterministic refine only). Args: path, optional language_hint, profile, debug_parser_payload.
vox_orchestrator_persistence_outbox_lifecycle	platform		Yes	Return the latest persistence outbox lifecycle health snapshot (queue size, replay successes/failures, and per-op replay failure counts).
vox_orchestrator_persistence_outbox_queue	platform		Yes	Inspect queued persistence outbox entries with optional lane filter, limit, and replay payload inclusion.
vox_orchestrator_start	ai		No	Probe embedded orchestrator + worker readiness. AgentFleet runs in-process when VOX_MCP_AGENT_FLEET is enabled (default); this tool does not spawn a separate process.
vox_orchestrator_status	ai		Yes	Get a full snapshot of the orchestrator state: agents, queues, and completed tasks.
vox_pause_agent	ai		No	Pause an agent queue (new work is held).
vox_pending_approvals	workflow		No	List human-in-the-loop tool-call approvals currently awaiting a decision.
vox_pending_questions	ai		No	List all questions waiting for my answer.
vox_pipeline_status	app		No	Get current compiler pipeline health and status (lexer, parser, typeck, codegen).
vox_plan	ai		No	Generate a Cursor-style structured task plan for a goal. Optionally writes PLAN.md to workspace root.
vox_plan_list_sessions	ai		No	List active and recent plan sessions.
vox_plan_node_upsert	workflow		No	Edit a not-yet-dispatched plan node’s description, or insert a new step (blocked while the plan is unapproved).
vox_plan_nodes_approve	workflow		No	Approve every blocked_on_approval node in a plan session; returns how many were released.
vox_plan_resume	ai		No	Resume a suspended or pending plan session.
vox_plan_status	ai		Yes	Plan session status via vox-orchestrator-d (ai.plan.status) for a session_id.
vox_plugin_catalog	interop		No	List plugins available in the Vox plugin catalog (marketplace index).
vox_plugin_info	interop		No	Show manifest details for one installed or catalog Vox plugin.
vox_plugin_install	interop		No	Install a Vox plugin by id from the catalog into the local plugins directory.
vox_plugin_list	interop		No	List installed Vox plugins with id, version, and discovered skill count.
vox_plugin_remove	interop		No	Remove an installed Vox plugin from the local plugins directory.
vox_poll_events	ai		No	Poll recent orchestrator events for all agents.
vox_populi_local_status	workflow		No	Return mens environment variables and the local mens registry file contents (CPU-first node records).
vox_preference_get	data		No	Get a user preference value by key from VoxDb.
vox_preference_list	data		No	List all user preferences, optionally filtered by a key prefix.
vox_preference_set	data		No	Set a user preference key to a value in VoxDb.
vox_project_init	platform		No	Scaffold Vox.toml, src/main.vox, .vox_modules/, or a skill markdown file — parity with vox init under the bound repo root (optional target_subdir).
vox_propose_skill	ai		No	Surface a mined recurring operation procedure as a non-blocking “save as skill?” suggestion in the NeedsYou inbox.
vox_publish_message	ai		No	Publish a message to the bulletin board for all agents to receive.
vox_questioning_pending	ai		No	Return open clarification session + unanswered question rows with MC options and belief_state_json for UI (requires VoxDb; same session_id as chat/plan).
vox_questioning_submit_answer	ai		No	Store the user’s answer to a pending Socrates clarification for an MCP session (requires VoxDb; pairs with socrates.questioning in chat/plan responses).
vox_questioning_sync_ssot	ai		No	Dual-write the information-theoretic questioning markdown SSOT into publication_manifests and search (requires VoxDb + workspace file).
vox_queue_status	ai		No	Returns the specific queue and tasks for an agent.
vox_rebalance	ai		No	Rebalances tasks dynamically across agents.
vox_record_cost	ai		No	Record a cost event from a Vox agent session token usage.
vox_redo	ai		No	Redo a previously undone operation.
vox_reorder_task	ai		No	Change the priority of a queued task.
vox_replan	ai		No	Replan via vox-orchestrator-d (ai.plan.replan): session_id + delta_hint; optional write_to_disk and mode.
vox_repo_catalog_list	platform		Yes	Resolve the explicit .vox/repositories.yaml catalog and return local plus remote repository descriptors.
vox_repo_catalog_refresh	platform		No	Re-resolve the explicit repo catalog and write a snapshot cache for the current workspace repository.
vox_repo_index_refresh	app		No	Refresh the on-disk repo index cache for the current workspace.
vox_repo_index_status	app		Yes	Return repo index cache status (bounded walk under repo root, .vox/cache/repos/...).
vox_repo_query_file	platform		Yes	Read a path across cataloged local repositories with per-repository path sandboxing.
vox_repo_query_history	platform		Yes	Read recent Git history per cataloged local repository, optionally filtered by path or substring.
vox_repo_query_text	platform		Yes	Search cataloged local repositories for text or regex matches and group results by repository_id.
vox_repo_status	platform		Yes	Discovered repository root, stable repository_id, stack markers, and sorted Cargo workspace member dirs (parity with vox repo status JSON).
vox_research_engine_config_set	data		No	Persist the research engine provider/lane config to the server’s Vox.toml and preferences.
vox_research_engine_probe	data		No	Probe one search provider (or all when provider is omitted) with a test query using server secrets.
vox_research_flag_misleading	data		No	Record that a research session’s answer misled the user (Axis Research surface).
vox_research_get	ai		No	MCP vox_research_get — fetch the completed research payload (answer + sources) for a session when available.
vox_research_publish_doc	data		No	Write a research session’s doc draft under docs/src/research on the server and store it as an artifact.
vox_research_run	ai		No	CLI vox research run and MCP vox_research_run — orchestrator run_research pipeline (web retrieval via vox-search, synthesis, judge when LLM configured).
vox_research_search	ai		No	MCP vox_research_search — full-text search over persisted research sessions, reports, and verified claims with confidence filtering.
vox_research_start	ai		No	MCP vox_research_start — creates a persisted research session and runs the same web-retrieval pipeline as vox_research_run in the background (returns session_id immediately).
vox_research_status	ai		No	MCP vox_research_status — read stage / progress for an async research session created by vox_research_start.
vox_resolve_approval	workflow		No	Resolve a pending human-in-the-loop approval (approve / modify / reject) by id.
vox_resolve_conflict	ai		No	Resolve a file conflict.
vox_resolve_feedback	ai		No	Resolve a pending clarification or doubt request with a specific action.
vox_resume_agent	ai		No	Resume a previously paused agent queue.
vox_retire_agent	ai		No	Retire an agent: release locks and affinity, unregister heartbeat, return remaining tasks.
vox_run_tests	app		No	Run cargo test for a specific crate, optionally filtered by test name.
vox_schola_submit	ai		No	Enqueue a background orchestrator task for Mens training intent; canonical execution remains vox mens train.
vox_scientia_assist_suggestions	data		No	Bounded Research-routed LLM assist — returns heuristic block plus optional JSON-only checklist/gap suggestions grounded on manifest scientia_evidence; never authoritative for novelty.
vox_scientia_discovery_ack	data		No	Acknowledge one discovery-inbox row so it leaves the unacknowledged list.
vox_scientia_publication_approve	data		No	Record one digest-bound approval for a scientia publication manifest.
vox_scientia_publication_arxiv_handoff_record	data		No	Append-only operator milestone for arXiv-assist workflow (publication_status_events with status arxiv_handoff:*; does not change manifest state). Stage published requires arxiv_id.
vox_scientia_publication_autofill	data		No	Plan (or apply) metadata autofill for a publication manifest from server workspace provenance.
vox_scientia_publication_claim_review	data		No	Record a human review decision for one claim of a publication.
vox_scientia_publication_decision_explain	data		No	CLI vox scientia publication-decision-explain; MCP vox_scientia_publication_decision_explain.
vox_scientia_publication_discovery_explain	data		No	Machine explanation JSON for one publication_id — discovery rank, completion report, evidence completeness, transform preview labels (machine_suggested / requires_human_review).
vox_scientia_publication_discovery_refresh_evidence	data		No	Merge live Socrates telemetry + JSON sidecars into a stored scientia manifest, refresh scientia_evidence (headings, discovery signals), upsert digest, append discovery_evidence_refreshed status (scientia only).
vox_scientia_publication_discovery_scan	data		No	Deterministic SCIENTIA discovery scan — rank publication_manifests rows by structured scientia_evidence signals (no LLM).
vox_scientia_publication_external_jobs_dead_letter	data		No	List external_submission_jobs in terminal failed status (not scheduled for retry), newest first.
vox_scientia_publication_external_jobs_due	data		No	List external_submission_jobs rows due for retry (queued, or retryable_failed with elapsed next_retry_at_ms) up to an optional limit.
vox_scientia_publication_external_jobs_replay	data		No	Requeue one dead-letter scholarly job (status failed → queued) by external_submission_jobs.id for the next publication_external_jobs_tick.
vox_scientia_publication_external_jobs_tick	data		No	Process one or more batches of due scholarly submit jobs (preflight dual-approval + digest match, lease row, submit using stored adapter). Optional lock_ttl_ms, lock_owner, iterations, interval_secs, max_runtime_secs, and jitter_secs between ticks.
vox_scientia_publication_external_pipeline_metrics	data		No	Read-only JSON rollup of scholarly external_submission_jobs (queue snapshot, in-window status view, percentiles, per-adapter terminal/retry ratios), attempts, snapshots, scholarly_submissions (incl. in-window view), and publication_attempts by channel. metrics_schema_version 2. Optional since_hours (default 168).
vox_scientia_publication_media_delete	data		No	Delete one publication media asset by publication_id + asset_ref.
vox_scientia_publication_media_list	data		No	List publication media assets for a scientia publication id.
vox_scientia_publication_media_upsert	data		No	Insert/update one publication media asset row for a scientia publication id.
vox_scientia_publication_nanopublish	data		No	Build and publish a nanopublication for a human-approved publication claim.
vox_scientia_publication_novelty_fetch	data		No	CLI vox scientia publication-novelty-fetch; MCP vox_scientia_publication_novelty_fetch.
vox_scientia_publication_novelty_happy_path	data		No	CLI vox scientia publication-novelty-happy-path; MCP vox_scientia_publication_novelty_happy_path.
vox_scientia_publication_preflight	data		No	Run publication_preflight checks on a stored scientia manifest; returns readiness_score and findings JSON (no writes). Optional with_worthiness adds heuristic worthiness rubric output from the repo default YAML.
vox_scientia_publication_prepare	data		No	Create/update a canonical scientia publication manifest and return its content digest. Optional scholarly_metadata (ScientificPublicationMetadata) is merged under metadata_json.scientific_publication.
vox_scientia_publication_publish	data		No	Publish a prepared publication id to selected channels (dry-run by default) and record publication_attempts. Live publish requires digest dual-approval, armed (orchestrator [news].publish_armed and/or VOX_NEWS_PUBLISH_ARMED), and respects orchestrator [news].dry_run merged with tool dry_run. Optional json=true returns a single-line compact JSON tool envelope.
vox_scientia_publication_retry_failed	data		No	Retry failed channels from the most recent publication attempt, optionally scoped to one channel. Optional json=true returns a single-line compact JSON tool envelope.
vox_scientia_publication_route_simulate	data		No	Simulate per-channel routing outcomes for a prepared publication id (policy + topic filters + template derivation).
vox_scientia_publication_scholarly_pipeline_run	data		No	Run scholarly pipeline on a stored manifest: preflight, dual-approval gate, optional venue staging export (output_dir + venue), then scholarly submit (unless dry_run). Optional json_compact returns a single-line JSON payload inside the tool result envelope (matches CLI —json).
vox_scientia_publication_scholarly_remote_status	data		No	Poll remote scholarly repository status for the latest stored submission (or a specific external_submission_id) using the adapter recorded in VoxDb.
vox_scientia_publication_scholarly_remote_status_sync_all	data		No	Poll remote scholarly status for every scholarly_submissions row on a publication (multi-deposit / multi-adapter); persists snapshots and best-effort continues per row.
vox_scientia_publication_scholarly_remote_status_sync_batch	data		No	Batch remote status poll: run sync-all logic for up to N publications ranked by recent scholarly_submissions activity (cron-friendly). Optional iterations, interval_secs, max_runtime_secs, and jitter_secs run repeated batches for a supervised worker.
vox_scientia_publication_scholarly_staging_export	data		No	Load a stored scientia publication manifest from VoxDb and write venue-scoped scholarly staging files (body.md, CITATION.cff, crossref_work.json, optional citations.json; zenodo.json for Zenodo; arxiv_handoff.json plus arxiv_bundle.tar.gz for arxiv-assist) under output_dir, then validate the tree.
vox_scientia_publication_status	data		No	Return manifest state, digest-bound approval count, discovery rank, manifest completion, transform preview stub, destination_readiness inside preflight, and scholarly submission rows.
vox_scientia_publication_submit_local	data		No	Submit an approved scientia publication through the scholarly adapter (optional adapter arg or VOX_SCHOLARLY_ADAPTER; default local_ledger), with external_submission_jobs ledger updates.
vox_scientia_worthiness_evaluate	data		No	Evaluate publication-worthiness.contract YAML (repo-relative or default) against WorthinessInputs metrics JSON; returns Publish / AskForEvidence / Abstain decision (no writes).
vox_search_callees	platform		No	Functions called by the given symbol(s) — forward call edges, constructor noise filtered (read-only).
vox_search_callers	platform		No	Functions that call the given symbol(s) — reverse call edges, constructor noise filtered (read-only).
vox_search_compare	platform		Yes	Diff two Vox Search structural-index manifests for node/edge/community delta (read-only).
vox_search_history	platform		No	First-parent repo history for agents (log through lineage, focus, forgotten, search, timeline, brief). Ingests at most 60 s per call into a cache under the git common dir; results carry complete/behind, and an incomplete result means call again.
vox_search_neighbors	platform		Yes	BFS neighbor expansion from seed node IDs (read-only; layered navigation).
vox_search_path	platform		Yes	Shortest path between two node IDs in the Vox Search structural index (read-only).
vox_search_rebuild	platform		No	Rebuild a Vox Search structural index (AST extract + cluster), snapshot the prior graph, and refresh freshness manifest (write).
vox_search_set_ttl	platform		No	Set the Vox Search corpora staleness TTL in days, writing the shared registry contract that the CLI and CI freshness gate also read (write).
vox_search_status	platform		Yes	Report Vox Search structural-index freshness (read-only; --strict fails when stale).
vox_search_structural	platform		No	Lexically search the Vox Search structural index on disk (read-only).
vox_secrets_doctor	platform		No	Returns a JSON snapshot of all managed secret resolution statuses. Agents should use this instead of probing environment variables directly.
vox_semantic_fs_discover	ai		Yes	Rank workspace paths for an intent via the semantic-fs bridge (read-only inventory search).
vox_session_cleanup	ai		No	Tick lifecycle and remove archived sessions.
vox_session_compact	ai		No	Replace a session’s history with a summary string.
vox_session_create	ai		No	Create a new persistent session for an agent.
vox_session_info	ai		No	Get detailed info about a specific session.
vox_session_list	ai		No	List all active sessions with state and token usage.
vox_session_reset	ai		No	Reset a session’s conversation history (keeps metadata).
vox_set_active_model	ai		No	Set sticky MCP chat model id for chat / inline / ghost (empty string clears).
vox_set_agent_budget	ai		No	Set custom VoxDB-powered budget limits for an agent.
vox_set_context	ai		No	Set a key-value pair in the shared orchestrator context store. Supports TTL.
vox_set_model	ai		No	Set per-agent model override in the orchestrator registry.
vox_skill_add	interop		No	Add a skill from a git URL or local path into the user .vox/skills root.
vox_skill_discover	interop		Yes	List bare SKILL.md skills found under the standard interop roots, with install status.
vox_skill_info	interop		No	Get detailed info on a specific skill by ID.
vox_skill_install	interop		No	Install a skill from a VoxSkillBundle JSON payload.
vox_skill_list	interop		No	List all installed skills.
vox_skill_parse	interop		No	Parse a SKILL.md and preview its manifest before installing.
vox_skill_remove	interop		No	Remove a user-installed skill directory (ownership-scoped to .vox/skills).
vox_skill_run	interop		No	Execute a skill script in the sandbox (parity with CLI vox skill run).
vox_skill_search	interop		No	Search installed skills by keyword.
vox_skill_uninstall	interop		No	Uninstall an installed skill by ID.
vox_skill_use	interop		No	Load the full SKILL.md body for an installed skill by id or name (tier-2 disclosure).
vox_snapshot_diff	ai		No	Show the file-level diff between two snapshots.
vox_snapshot_list	ai		No	List recent file snapshots for an agent.
vox_snapshot_restore	ai		No	Restore files to a previous snapshot state.
vox_spawn_agent	ai		No	Spawn a new orchestrator agent queue (optional dynamic/transient agent for auto-retire).
vox_speech_to_code	ai		No	Orchestrated speech-to-code: transcribe workspace audio (Oratio) or accept prompt text, then vox_generate_code with shared session_id; optional route_mode, emit_trace_path JSONL for MENS.
vox_submit_task	ai		No	Submit a new task to the orchestrator. Routes to the best agent by file affinity.
vox_suggest_model	ai		No	Suggest the best model for a task category string (codegen, review, etc.).
vox_task_status	ai		Yes	Get the current status of a specific task by ID.
vox_test_all	app		No	Run cargo test for the entire workspace.
vox_test_decision	ai		Yes	Get the Testing Decision Engine’s requirement classification (Required/Recommended/Deferred/Skip) for a task by ID.
vox_tool_search	ai		Yes	Search the MCP tool registry by keyword and return matching tools (name, description, input schema) for progressive tool disclosure.
vox_transfer_file	ai		No	Transfer ownership of a file to another agent.
vox_trust_override	ai		No	Overrides the trust tier for the agent context.
vox_undo	ai		No	Undo the last operation or a specific operation by ID.
vox_validate_file	app		No	Validate a .vox file using the full compiler pipeline (lexer → parser → typeck → HIR).
vox_validate_source	app		No	Validate Vox source code passed as a string (no filesystem read). Returns LSP-style diagnostics with stable codes and structured autofix suggestions for the iterative AI-assistant repair loop.
vox_validate_vuv	ai		Yes	Validate Vox/VUV source against the compile-time GUI guarantees (contrast, layer-occlusion, a11y, structural web-IR) without writing files.
vox_vcs_status	ai		No	Get unified VCS status.
vox_verify_task_claims	ai		No	Check receipt ids an agent claims against the orchestrator’s receipt ledger; returns which are valid, fabricated (no ledger entry) or unverified (tag mismatch). Read-only.
vox_visual_rag_query	ai		No	NOT IMPLEMENTED: multimodal RAG is planned but not wired up; this tool always returns an error. Use vox_search or memory retrieval tools instead.
vox_visus_audit	ai		No	Performs an automated GUI visual audit (screenshot, AXTree, overlaps).
vox_visus_baseline	ai		No	Manages golden visual baselines for regression testing.
vox_workspace_create	ai		No	Create an isolated workspace for an agent.
vox_workspace_mcp_refresh	interop		No	Reloads/refreshes the workspace MCP registry and scan config.
vox_workspace_merge	ai		No	Merge an agent’s workspace changes back to main.
vox_workspace_modules	app		Yes	Returns a list of all .vox files in the workspace repository.
vox_workspace_status	ai		No	Show files modified in an agent’s workspace.