| name | kg-analyze |
| skill_type | skill |
| description | Runs residual ops/structural analysis across the KG — inspect, enrichment coverage, process write-back (Camunda/ARIS), workload placement plan, infra sweep, security scan. Use for structural/ops analysis — "inspect this subgraph", "placement plan", "security scan". (Code→kg-code, research→kg-research, eval→kg-evaluate, Q&A→kg-explain.) |
| license | MIT |
| tags | ["graph-os","analyze"] |
| tier | core |
| metadata | {"author":"Genius","version":"0.1.0"} |
kg-analyze
Condensed intent-surface note (Seam 8). Under the default intent surface (MCP_TOOL_MODE=intent), graph_analyze is held back from the default tool list (nothing removed — REST + _execute_tool still reach it exactly as documented below). Two ways to use this skill unchanged: (1) load_tools(tools=["graph_analyze"]) once per session (as below), then proceed exactly as documented; or (2) call the ask intent verb with the same natural-language request — the resolver routes to graph_analyze for you and returns the result plus a routing justification. Set MCP_TOOL_MODE=condensed/verbose/both to expose the granular tools eagerly instead.
graph_analyze handles the ops/structural analysis actions: inspect, enrichment_coverage, process_writeback (push KG intelligence into Camunda/ARIS; target=camunda|aris|both), placement_plan (workload placement, KG-2.9), infra_sweep, security_scan. Codebase, research, evaluation and Q&A intents are routed to their own verbs (kg-code, kg-research, kg-evaluate, kg-explain).
Invoke
- MCP:
load_tools(tools=["graph_analyze"]), then graph_analyze(action="inspect", query="...").
- REST twin:
POST /graph/analyze with {"action": "placement_plan", "query": "..."}.
Example
graph_analyze(action="security_scan", query="services/")
Delegation
If graph-os is reachable, offload composite multi-step work via graph_orchestrate (execute_agent / execute_workflow) instead of hand-running the steps — let the local LLM + Loop engine do it, and resolve only the exceptions.