slaunt
DOCUMENTATION

Tools overview

Slaunt's tools fall into a few groups. The lifecycle and Context tools are the ones nearly every agent uses on every turn; the rest you reach for when the task needs them.

GroupUse it toTools
LifecycleReport threads and runs to Quaerothread_begin, run_begin, run_step, run_end, thread_close
Slaunt ContextLook up and record what happenedccr_search, ccr_recent, ccr_log_event, ccr_issue_log, ccr_event_*
TasksTrack work that's still opentasks_*
WorkspaceCheck identity and load context in one callidentity_context, aki_workspace_context, workspace_people_list
RoomRead and organize shared filesroom_*
GitHubVerify against repos, PRs, CIgithub_*
Hooks and captureUsed by client hooks and bridges, not the modelcodex_hook_event, claude_hook_event, chat_*, tool_authorize*, threads_*
Provider toolsSlack, Notion and other connected providerse.g. slack.*

Tools that only appear sometimes

ToolsListed when
claude_hook_eventThe connection is an authenticated Claude Code client
chat_capture, chat_hook_event, chat_events_listChats is enabled on the server
threads_list, threads_readThreads agent access is enabled (HTTPS only). chat_events_list is hidden in this mode.
Provider toolsThe workspace has connected providers

Spawning, messaging or killing agents isn't possible over MCP; use Quaero or KillBar.

Result format

Results carry a readable summary for the model and the full data for code:

{
  "content": [{ "type": "text", "text": "Found 1 semantic matches for \"auth deploy\".\n- change: Deployed auth-service v2.3.1 to production. score=0.620 tags=[deploy] @ 2026-11-12T18:04:11.000Z (id=...)" }],
  "structuredContent": { "matches": [{ "score": 0.62, "record": { "...": "..." } }] }
}

Response shapes in the tool sections describe structuredContent. The optional org_id/project arguments on many tools are only a cross-check against your token; leave them out.