2dph/skills/brain/SKILL.md

name, description

name	description
brain	Deduction search over the 2dph brain (Ladybug graph: ops corpus, portfolio, ssh hosts) with bin/brain/search.go instead of reading files or grepping repos. Use whenever a question starts with "where is", "what runs on", "which file describes", "who is", "how is X done", before opening any documentation.

brain — deduction over facts and info

One embedded Ladybug graph (var/kb.lbug, read-only when queried) holding two roots:

• facts — assertions backed by ≥2 independent sources (docker ps × compose × ssh-config × docs), confidence: confirmed.
• info — descriptive/narrative leafs, searchable, never asserted.

Search = deduction: facts root first, info root second, web-search as the second independent source when local roots cannot confirm. An answer is confirmed only if it comes off the facts root; anything else is (not confirmed).

bin/brain/search.go "Matrix federation"                # pointers + snippets, YAML
bin/brain/search.go "onlyoffice postgres" --root facts # restrict to confirmed
bin/brain/search.go "where is cs-lexicon" --json | yq '.[].ref'
bin/brain/get.go <id> --body                           # full chunk only when needed
bin/brain/stats.go                                     # index health
bin/brain/eval.go                                      # recall@5 >= 0.95 gate (Go; Python bin/kb/eval is CI fallback)

bin/kb/search is a deprecated wrapper. --hop errors (File/FROM_FILE edges are not wired yet); do not treat it as a graph walk.

Rules

• Search before you read. Never grep a repo for a concept the graph covers.
• --root facts returns only confirmed evidence-linked answers. Default shows facts first, then info leafs clearly marked (not confirmed).
• If there is no facts hit, bin/brain/search.go consults SearXNG and adds a web block (kept apart from graph hits). throttled / skipped / refused are not evidence of absence. --root facts|info and --no-web skip the web.
• If recall looks wrong, run bin/brain/eval.go; it gates control questions and should stay at or above 95% recall@5.
• Agents: GET /openapi.json and POST /mcp on bin/brain/serve.go (same handlers; tool names match paths search/get/stats/audit). Generated list: tools.md.
• Never report an unconfirmed single-source local answer as fact.