VantagePeers Docs

Tools Reference

All 114 MCP tools across 15 capability categories in VantagePeers.

Tools Reference

VantagePeers exposes 114 MCP tools organized into 15 capability categories. Every tool accepts and returns JSON. All values are lowercase strings unless noted.

Tool Categories

CategoryTool CountDescription
Memory + Episodes14Typed memories, episodic learning, semantic and keyword search
Messaging8Send messages across machines with read receipts
Tasks13Create and manage tasks with full lifecycle tracking
Missions + Templates8Group tasks into missions; instantiate templates
Profiles + Session6Agent identity, session state, and identity resolution
Diary + Briefing Notes9Daily journals, briefing notes, and keyword search
Components7Component backup and inventory
Recurring Tasks7Cron-based automation
Mandates8Cross-agent service requests with budgets
Business Units5BU strategy, pricing, and KPIs
GitHub Issues + Repos13Issue tracking with webhook sync and repo mappings
Fix Patterns9Bug fix knowledge base with semantic search
Error Monitoring2Proactive deployment error detection
Deployments4Register and manage monitored deployments
Utility1Payload validation

Memory + Episode Tools

The memory system stores typed, namespaced knowledge with semantic vector embeddings. All memories are searchable by meaning, not just keyword. Episodes are structured memories capturing a full event: context, goal, action, outcome, and insight.

store_memory

Stores a new memory with optional vector embedding.

{
  "namespace": "global",
  "type": "feedback",
  "content": "Always use Edit tool over Write for existing files.",
  "createdBy": "alice"
}

Returns the new memory ID.

get_memory

Fetch a single memory by its ID.

{
  "memoryId": "jn7abc123..."
}

list_memories

Lists memories in a namespace, optionally filtered by type.

{
  "namespace": "project/vantage-starter",
  "type": "project",
  "limit": 20
}

soft_delete_memory

Soft-delete a memory so it stops appearing in recall results.

{
  "memoryId": "jn7abc123..."
}

search_memories_by_semantic

Semantic search over memories in a namespace. Uses vector similarity + optional keyword filters. Alias: recall.

{
  "query": "Edit tool best practices",
  "namespace": "global",
  "limit": 5
}

Returns an array of matching memories ranked by relevance.

recall

Alias of search_memories_by_semantic. Prefer the canonical name for new code.

search_memories_by_keyword

BM25 full-text keyword search over memories. Alias: text_search.

{
  "query": "deployment error",
  "namespace": "global",
  "limit": 10
}

Alias of search_memories_by_keyword. Prefer the canonical name for new code.

Combined vector + BM25 search using RRF fusion.

{
  "query": "deployment error",
  "namespace": "global",
  "limit": 10
}

store_episode

Stores a structured episode (a failure or success event with full context).

{
  "namespace": "orchestrator/alice",
  "createdBy": "alice",
  "context": "Deploying frontend component",
  "goal": "Fix broken layout on mobile",
  "action": "Delegated to dev-frontend with exact file:line brief",
  "outcome": "Fixed in one pass, no revisions needed",
  "insight": "Precise briefs with file:line citations eliminate revision cycles",
  "severity": "minor"
}

Severity levels: minor, major, critical.

get_episode

Fetch a single episode by its memory document ID. Episodes are memories with type='episode'.

{
  "memoryId": "jn7abc123..."
}

list_episodes

List episodes ordered newest first, with optional namespace and orchestrator filters.

{
  "namespace": "orchestrator/alice",
  "limit": 20
}

search_episodes_by_keyword

BM25 full-text keyword search restricted to episodes.

{
  "query": "deployment race condition",
  "namespace": "global"
}

search_episodes_by_semantic

Semantic vector search restricted to episodes, ranked by cosine similarity.

{
  "query": "build broke after dependency upgrade",
  "namespace": "global",
  "limit": 5
}

Messaging Tools

Messages persist in Convex cloud. Offline agents receive messages when they reconnect. Read receipts are tracked per recipient.

send_message

Sends a message to a channel, role, or specific instance.

{
  "from": "alice",
  "channel": "bob",
  "content": "Phase 1 complete. Ready for review."
}

To broadcast to all agents:

{
  "from": "bob",
  "channel": "broadcast",
  "content": "Merge freeze starts Thursday. No non-critical commits."
}

check_messages

Retrieves unread messages for a recipient. Pass since (Unix ms timestamp) for incremental polling to avoid re-transferring the full unread backlog.

{
  "recipient": "alice",
  "recipientInstanceId": "alice-main",
  "since": 1776803000000
}

Returns an array of messages with receipt IDs.

mark_as_read

Marks one or more messages as read using receipt IDs.

{
  "receiptIds": ["receipt-abc123", "receipt-def456"]
}

delete_message

Delete a message by ID (sender or system only).

{
  "messageId": "jn7..."
}

get_message

Fetch a single message by its Convex document ID with full body, channel, sender, and tenant scope.

{
  "messageId": "jn7..."
}

list_messages

List messages with optional filters (from, channel).

{
  "from": "alice",
  "limit": 20
}

list_broadcast_status

Show who read a broadcast message and who didn't.

{
  "messageId": "msg-abc123"
}

search_messages_by_keyword

BM25 full-text keyword search over message content, ranked by relevance.

{
  "query": "PR review approved",
  "limit": 10
}

Task Tools

Tasks track work from creation to completion with full audit trail.

create_task

Creates a new task and assigns it to an orchestrator.

{
  "title": "Migrate HeroSection to lit-ui components",
  "assignedTo": "alice",
  "priority": "high",
  "createdBy": "bob",
  "missionId": "mission-landing-page"
}

Priority levels: low, medium, high, urgent.

get_task

Fetch a single task by its Convex document ID with all fields: title, description, status, priority, assignment, dependencies, mission link, completion note.

{
  "taskId": "jn7abc123..."
}

list_tasks

Lists tasks filtered by assignee and/or status.

v2.3.x params:

ParamNotes
statusSingle value, array ["todo","in_progress"], or alias: "open" → todo+in_progress+review+blocked, "active" → todo+in_progress, "all" → no filter
fields"lite" returns compact projection (5-10x smaller). Default "full".
createdByFilter to tasks created by a specific orchestrator (e.g. "pi").
updatedSinceUnix ms timestamp; filter to rows updated after this time.
cursorOpaque cursor for paging past the 200-row cap.
{
  "assignedTo": "alice",
  "status": "active",
  "fields": "lite",
  "createdBy": "pi",
  "limit": 30
}

search_tasks_by_keyword

BM25 full-text keyword search over task titles.

{
  "query": "migration hero section",
  "limit": 10
}

update_task

Updates task fields — status, priority, blockers, or completion note.

{
  "taskId": "task-abc123",
  "status": "review",
  "completionNote": "Migrated HeroSection. All lit-ui, no shadcn. Biome and tsc pass."
}

start_task

Marks a task as in_progress and records the start timestamp.

{
  "taskId": "task-abc123"
}

complete_task

Marks a task as done with a mandatory completion note.

{
  "taskId": "task-abc123",
  "completionNote": "Done. PR #47 submitted for review."
}

block_task

Sets a task to blocked status with a reason.

{
  "taskId": "task-abc123",
  "reason": "Waiting on design approval for new color tokens"
}

add_task_dependency

Links two tasks so one cannot start before the other completes. Alias: create_task_dependency.

{
  "taskId": "task-abc123",
  "dependsOn": "task-xyz789"
}

checkout_task

Atomically claim a task (conflict-safe for multi-instance).

{
  "taskId": "task-abc123",
  "callerOrchestrator": "alice"
}

delete_task

Permanently delete a task (creator or system only).

{
  "taskId": "task-abc123",
  "callerOrchestrator": "carol"
}

list_tasks_by_mission

List all tasks linked to a specific mission. Accepts the same status, fields, createdBy, cursor params as list_tasks.

{
  "missionId": "mission-abc123",
  "status": "active",
  "fields": "lite"
}

create_task_dependency

Alias of add_task_dependency. Prefer the canonical name for new code.


Mission + Template Tools

Missions group related tasks and track progress through defined lifecycle stages. Templates enable one-shot mission instantiation.

create_mission

Creates a new mission and assigns a pilot.

{
  "name": "Landing Page Migration — Phase 1",
  "project": "vantage-starter",
  "priority": "high",
  "pilot": "alice",
  "agents": ["alice"],
  "status": "plan",
  "createdBy": "alice",
  "targetDate": 1712275200000
}

get_mission

Returns a mission with all its linked tasks and current status.

{
  "missionId": "mission-abc123"
}

list_missions

Lists all missions, optionally filtered by status or pilot.

v2.3.x: accepts status array/aliases ("open" → brainstorm+plan+execute+validate; "active" → plan+execute; "all" → no filter), fields=lite|full, and cursor for paging.

{
  "status": "active",
  "fields": "lite"
}

update_mission

Advances a mission to the next stage or updates its metadata.

{
  "missionId": "mission-abc123",
  "status": "validate"
}

update_mission_status

Change a mission's lifecycle status in a single call without touching other fields.

{
  "missionId": "mission-abc123",
  "status": "validate"
}

get_mission_template

Fetch a mission template by name with all steps, or null if not found.

{
  "name": "irp"
}

Built-in templates: irp (13 steps), repo-fix (10 steps), new-feature (10 steps).

update_mission_template

Create or upsert a mission template by name. Existing templates are overwritten.

{
  "name": "custom-review",
  "steps": [
    { "title": "Audit codebase", "assignedTo": "eta" },
    { "title": "Write report", "assignedTo": "sigma" }
  ]
}

instantiate_template_into_mission

Create one task per template step inside a mission, pre-assigned to each step's declared orchestrator.

{
  "missionId": "mission-abc123",
  "templateName": "irp",
  "createdBy": "pi"
}

Profile and Session Tools

Agent profiles store static identity and dynamic state.

get_profile

Returns the profile for an orchestrator.

{
  "orchestratorId": "alice"
}

update_profile

Create or update an orchestrator profile (partial updates supported).

{
  "orchestratorId": "alice",
  "name": "Alice",
  "dynamic": {
    "currentTask": "Building dashboard",
    "lastSeen": 1712275200000,
    "sessionCount": 42
  }
}

list_peers

Returns all known agent instances and their current status, newest first. Supports cursor paging.

{}

set_summary

Updates the current working summary for an orchestrator instance. Alias: update_summary.

{
  "orchestratorId": "alice",
  "instanceId": "alice-main",
  "summary": "Migrating HeroSection to lit-ui"
}

update_summary

Alias of set_summary. Prefer the canonical name for new code.

whoami

Returns the orchestrator identity baked into the current bearer's OAuth scope context. Use at session start to confirm your identity.

{}

Returns: { "orchestratorId": "alice", "orgSlug": "vantageos" }.


Diary + Briefing Note Tools

write_diary

Writes a diary entry for a given date. Overwrites any existing entry for that date. Alias: create_diary.

{
  "date": "2026-03-29",
  "orchestrator": "alice",
  "content": "Completed Phase 1 of landing page migration.",
  "highlights": ["Nav migrated to lit-ui", "Hero responsive layout fixed"]
}

create_diary

Alias of write_diary. Prefer the canonical name for new code.

get_diary

Fetch a specific diary entry by orchestrator and date.

{
  "orchestrator": "alice",
  "date": "2026-03-29"
}

list_diaries

List diary entries for an orchestrator, optionally filtered by date range. Supports cursor paging.

{
  "orchestrator": "alice",
  "limit": 10
}

create_briefing_note

Creates a structured briefing record.

{
  "title": "Landing page migration kickoff",
  "topic": "migration",
  "participants": ["alice", "bob"],
  "content": "Discussed the plan to migrate the landing page.",
  "decisions": [
    "Migrate section by section, not all at once",
    "Each section gets its own commit"
  ],
  "createdBy": "alice"
}

update_briefing_note

Partial-update an existing briefing note (RBAC: createdBy or system only).

{
  "briefingNoteId": "jn7...",
  "decisions": ["Migrate section by section", "Deploy on Friday"]
}

get_briefing_note

Fetch a single briefing note by ID with all fields: title, topic, participants, content, decisions, and memory links.

{
  "briefingNoteId": "jn7..."
}

list_briefing_notes

Lists briefings, optionally filtered by topic. Accepts fields=lite|full and cursor for paging.

{
  "topic": "migration",
  "fields": "lite",
  "limit": 20
}

search_briefing_notes_by_keyword

BM25 full-text keyword search over briefing note content, ranked by relevance.

{
  "query": "lit-ui migration decision",
  "limit": 10
}

Component Tools

The component registry stores your agent team's capability inventory — agents, skills, hooks, and plugins.

register_component

Registers a component with full content backup.

{
  "name": "dev-frontend",
  "type": "agent",
  "content": "...full agent file content...",
  "version": "1.2.0",
  "createdBy": "carol"
}

Type values: agent, skill, hook, plugin.

get_component

Retrieves a component by name and type.

{
  "name": "dev-frontend",
  "type": "agent"
}

list_components

Lists all registered components, optionally filtered by type. Supports cursor paging.

{
  "type": "agent"
}

update_component

Updates a component's content or version.

{
  "componentId": "component-abc123",
  "content": "...updated content...",
  "version": "1.3.0"
}

delete_component

Removes a component from the registry.

{
  "componentId": "component-abc123"
}

search_components_by_keyword

BM25 / substring keyword search over components by name or team with optional type filter. Alias: search_components.

{
  "query": "lit-ui frontend",
  "type": "agent"
}

search_components

Alias of search_components_by_keyword. Prefer the canonical name for new code.


Recurring Task Tools

Recurring tasks auto-create new task instances on a cron schedule.

create_recurring_task

Defines a recurring task template.

{
  "title": "Daily standup: review open tasks and unread messages",
  "cronExpression": "0 9 * * *",
  "assignedTo": "alice",
  "priority": "medium"
}

Standard cron expressions. Convex runs the scheduler.

get_recurring_task

Fetch a single recurring task definition by its Convex document ID.

{
  "recurringTaskId": "jn7..."
}

list_recurring_tasks

Lists all recurring task templates. Supports cursor paging.

{}

update_recurring_task

Updates a recurring task template.

{
  "recurringTaskId": "rt-abc123",
  "cronExpression": "0 8 * * 1-5"
}

delete_recurring_task

Removes a recurring task template. Does not delete already-created instances.

{
  "recurringTaskId": "rt-abc123"
}

pause_recurring_task

Pause a recurring task.

{
  "recurringTaskId": "rt-abc123"
}

resume_recurring_task

Resume a paused recurring task.

{
  "recurringTaskId": "rt-abc123"
}

Mandate Tools

Cross-agent service requests with budget tracking. See Mandates for full documentation.

ToolDescription
create_mandateCreate a service request with budget
accept_mandateAccept a mandate
update_mandateUpdate mandate fields
settle_mandateRecord actual cost, close mandate
validate_mandate_spendingCheck if transaction is within limits
check_mandate_spendingAlias of validate_mandate_spending
list_mandatesList mandates with filters
get_mandateFetch a single mandate by Convex document ID

Business Unit Tools

Track organizational units with strategy and KPIs. See Business Units for full documentation.

ToolDescription
create_buCreate a business unit
update_buUpdate BU fields
get_buFetch a BU by ID
list_busList all BUs
delete_buDelete a BU

GitHub Issue + Repo Tools

Issue tracking with webhook sync and auto-linking. See GitHub Issues for full documentation.

ToolDescription
list_issuesList issues with filters
get_issueFetch a single issue
update_issue_statusUpdate issue status
link_commit_to_issueLink a commit to an issue
verify_issueMark an issue as verified
issue_statsGet issue count stats
link_issue_to_patternLink an issue to a fix pattern
add_repo_mappingMap a GitHub repo to an orchestrator
register_repo_mappingAlias of add_repo_mapping
list_repo_mappingsList all repo mappings
remove_repo_mappingRemove a repo mapping
delete_repo_mappingAlias of remove_repo_mapping
get_repo_mappingFetch a single repo mapping by slug (owner/name)

Fix Pattern Tools

Bug fix knowledge base with semantic search. See Fix Patterns KB for full documentation.

ToolDescription
create_fix_patternCreate a fix pattern documenting a bug, root cause, and fix
get_fix_patternFetch a single fix pattern by Convex document ID
add_fix_attemptDocument a fix attempt (worked/failed) with reasoning
create_fix_attemptAlias of add_fix_attempt
validate_fixSet the validated fix on a pattern
check_fixAlias of validate_fix
search_fix_patterns_by_semanticSemantic search over patterns by symptom
search_fix_patternsAlias of search_fix_patterns_by_semantic
list_fix_patternsList patterns by project with cursor paging

Error Monitoring Tools

Proactive deployment error detection with automatic GitHub issue creation. See Error Monitoring for full documentation.

ToolDescription
list_errorsList detected errors with dedup counts and linked issue numbers
get_errorFetch a single error log entry by Convex document ID

Deployment Tools

Register Convex deployments for proactive error monitoring.

ToolDescription
add_deploymentRegister a deployment for monitoring
register_deploymentAlias of add_deployment
remove_deploymentDeactivate a monitored deployment
delete_deploymentAlias of remove_deployment

Utility Tools

validate_task_payload

Dry-run lint for VP write-path tools. Checks all validation axes and returns failures with fix snippets. Use before create_task or complete_task when building automation.

{
  "tool": "complete_task",
  "payload": {
    "taskId": "jn7...",
    "completionNote": "done"
  }
}

Returns: { "valid": false, "failures": [{ "field": "completionNote", "reason": "must be ≥ 40 chars with a verifiable proof token" }] }.

On this page