← ClaudeAtlas

edgepilot-researchlisted

Route anonymous public strategy discovery and reproducible historical backtests through the local EdgePilot Research Runtime. It has no accounts, credentials, paper, demo, live trading, or order execution.
Rivendell-investment/edgepilot-research-claude · ★ 0 · Testing & QA · score 60
Install: claude install-skill Rivendell-investment/edgepilot-research-claude
# EdgePilot Research router The Node Ready Bridge always exposes `edgepilot_runtime_status`, `edgepilot_runtime_start`, `edgepilot_runtime_update` and `edgepilot_runtime_repair`, even before Runtime exists. The bridge automatically prepares the release-bound Runtime before first business use; never treat an older compatible Runtime as ready. When Runtime is ready, use the five Host meta tools: 1. `edgepilot_connection_list` 2. `edgepilot_tool_search` 3. `edgepilot_tool_get` 4. `edgepilot_tool_execute` 5. `edgepilot_result_present` Search for an operation, fetch its exact current descriptor, then execute with the returned `schema_revision`. Do not invent dynamic tools, copy operation schemas into this plugin or infer missing benchmark facts. Preserve exact strategy version, package digest, configuration-schema digest, dataset provenance and Runtime identity in reproducible results. Route one user outcome at a time. Use `edgepilot_connection_list` for current anonymous catalog availability. Search unknown capabilities once in concise English, batch Get for the exact contracts, and batch Execute only for independent calls. A returned identifier or digest starts a later dependency round. Present only an execute-minted `result_ref`. The normal research route is catalog search/recommend, exact inspect/version list, install, configuration resolve, backtest start, durable job status and result get. Runtime workflow hints are navigation, not execution authority. Never repeat a st