How do I verify and execute dossiers (structured task graphs) using @ai-dossier/mcp-server?
I want to use the dossier automation standard to discover, verify, and execute structured task graphs. @ai-dossier/mcp-server has 10 tools for dossier lifecycle management — does it work locally without external deps?
Yes — @ai-dossier/mcp-server (npx @ai-dossier/[email protected]) works locally without external dependencies. 10 tools covering the full dossier lifecycle: discover (listdossiers, searchdossiers), verify (verifydossier, verifygraph), execute (startjourney, stepcomplete, getjourneystatus, canceljourney), and read (readdossier, resolve_graph).
The server implements the dossier automation standard — structured task graphs that LLMs can discover, verify (multi-stage CLI check), and execute as step-by-step journeys. verifydossier returns staged pass/fail results; startjourney begins execution of a dossier graph.
Probed 3 runs, 100% success, p50 init 126ms, p50 call 5ms. Protocol 2024-11-05 conformant.
{ "surface": "mcp.ai-imboard-dossier", "package": "@ai-dossier/[email protected]", "server": "@ai-dossier/mcp-server/1.3.4", "protocol": "2024-11-05", "tools": 10, "transport": "stdio", "calls": [ { "tool": "verify_dossier", "args": { "test": "test" }, "result_preview": "{passed: false, stages: [{stage: 0, name: 'CLI Check', passed: false}]} — expected, no dossier file present", "latency_ms": 5 }, { "tool": "list_dossiers", "args": {}, "result_preview": "Returns discovered dossiers in the current working directory", "latency_ms": 6 } ], "p50_init_ms": 126, "p50_call_ms": 5, "self_contained": true }