# fthsrbst/ai-hub

Source: https://glimind.com/server/glama%2Ffthsrbst%2Fai-hub
Registry id: `glama/fthsrbst/ai-hub`
Status: unknown

## Tools (75)

- `agent_active`: Lists agents checked in on a project with heartbeat status.
- `agent_checkin`: Announces which device and branch an agent is working on.
- `agent_checkout`: Closes a presence record as done or abandoned.
- `agent_find`: Finds agents with a specific capability, optionally filtered.
- `agent_handoff`: Builds structured context-handoff package for delegation.
- `agent_inbox`: Returns unread messages for an agent.
- `agent_list`: Lists all registered agents with capabilities and status.
- `agent_message_send`: Sends a message to another agent or broadcasts.
- `agent_purge_stale`: Closes every stale active presence record.
- `agent_register`: Registers or updates an agent's capabilities and concurrency.
- `audit_list`: Reads redacted node-local request audit events.
- `audit_verify`: Verifies node-local audit hash chain integrity.
- `compact_project`: Triggers knowledge compaction summarizing sessions and decisions.
- `context_get`: Combines project map, session, memories and RAG evidence with intent-aware ordering.
- `event_log`: Returns recent hub events for debugging or monitoring.
- `graph_neighbors`: Pages through immediate neighbors with edge direction and confidence.
- `graph_node`: Resolves a node and its degree without loading the whole graph.
- `hygiene_report`: Reports memory quality including duplicates and contradictions.
- `hygiene_run`: Runs automated hygiene pass archiving stale memories.
- `integrity_check`: Read-only audit for unknown references, conflicts and orphans.
- `job_enqueue`: Adds async job to worker queue.
- `job_status`: Checks status of a specific job or lists recent jobs.
- `knowledge_transfer`: Finds applicable knowledge from other projects by tag overlap.
- `local_llm`: Runs a prompt against a registered local model.
- `machine_register`: Registers or updates a machine running local AI services.
- `machine_status`: Returns live status and loaded models for registered services.
- `media_generate`: Generates image, video, audio or 3D asset via ComfyUI.
- `memory_consolidate`: Merges duplicate memories into one target, tombstoning sources.
- `memory_delete`: Removes an incorrect or no longer valid memory.
- `memory_relation_add`: Creates a directional temporal edge between two memories.
- `memory_relation_delete`: Deletes a provably incorrect edge.
- `memory_relation_list`: Inspects typed edges by memory, relation type, or timestamp.
- `memory_relation_update`: Updates confidence, temporal validity, source or metadata on an edge.
- `memory_save`: Stores a durable fact, decision, preference, how-to or project context.
- `memory_search`: Hybrid keyword and semantic search across stored memories.
- `memory_update`: Corrects a memory that has gone stale or been proven wrong.
- `message_mark_all_read`: Marks all direct messages and broadcasts as read.
- `message_mark_read`: Marks a single message as read.
- `message_unread_count`: Returns unread message count for an agent.
- `metrics_overview`: Returns system metrics including uptime and latency percentiles.
- `profile_get`: Returns the canonical professional profile and metadata.
- `profile_update`: Replaces the canonical profile document.
- `project_add_decision`: Appends one line to the project decision history.
- `project_delete`: Permanently deletes a project map with tombstone propagation.
- `project_detach_references`: Rewrites pseudo-project references to global scope.
- `project_get`: Full project context including summary, stack, decisions and code map.
- `project_lessons`: Returns aggregated lessons learned from task feedback.
- `project_list`: Lists every project map with its status.
- `project_migrate_references`: Atomically moves references from stale name onto canonical project.
- `project_update`: Creates or merges into a project map.
- `prompt_get`: Returns a role's system prompt prefixed with engineering core.
- `prompt_list`: Lists role-based system prompts on the hub.
- `rag_add`: Indexes a document or note into the RAG archive with chunking.
- `rag_search`: Hybrid search across indexed documents returning source-referenced chunks.
- `recall`: Returns memories and document chunks relevant to one message or task.
- `recall_feedback`: Reports whether recalled evidence was helpful, noisy or missing.
- `session_log`: Records what was finished, left open and next step.
- `session_recent`: Returns the most recent session logs.
- `skill_list`: Lists the agent skills stored on the hub.
- `skill_save`: Creates or updates a skill in SKILL.md format.
- `task_claim`: Claims a specific task or next available from project queue.
- `task_complete`: Marks a task done with optional structured result.
- `task_create`: Creates a new task for agent-to-agent work delegation.
- `task_feedback`: Records feedback for a completed task with lessons learned.
- `task_list`: Lists tasks with optional filters by project, status or tags.
- `task_queue`: Returns next actionable tasks for a project ordered by priority.
- `task_update`: Updates task status, priority or other fields.
- `vector_projection_flush`: Attempts one bounded delivery batch now.
- `vector_projection_rebuild`: Queues every authoritative vector for idempotent redelivery.
- `vector_projection_status`: Reports vector backend readiness and outbox depth.
- `vector_projection_verify`: Compares sqlite-vec counts against external generation.
- `webhook_list`: Lists all registered webhooks with their status.
- `webhook_register`: Registers HTTP endpoint for hub events with HMAC signing.
- `webhook_remove`: Removes a registered webhook by UID.
- `workflow_list`: Lists available ComfyUI generation workflows.

Measured by Glimind (https://glimind.com): MCP handshake + tool list only, no tool is ever called. CC BY 4.0.