Run the Knowledge Brain (citation-governed retrieval over meeting transcripts: ask/cite/prep/remediate) or its verbatim-or-refuse integrity loop. Use for cited answers, review-to-remediation, or KB drift repair.
Scanned 8/31/2026
Install to Claude Code
npx -y skills add sophonfinance-wq/finance-automation-portfolio --skill brain --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Brain?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/sophonfinance-wq-brain)More formats (shields.io, HTML) on the badges page.
---
name: brain
description: "Run the Knowledge Brain (citation-governed retrieval over meeting transcripts: ask/cite/prep/remediate) or its verbatim-or-refuse integrity loop. Use for cited answers, review-to-remediation, or KB drift repair."
---
# Knowledge Brain Engine
Package `brain_engine` in `knowledge-brain-engine/`. Stdlib TF-IDF, no network. Exit codes: 0 sourced, 2 usage error, 3 refusal.
## Commands (run from `knowledge-brain-engine/`)
```bash
python -m brain_engine ask "how do we handle FX on ACB?"
python -m brain_engine --cite "distributions draw exempt surplus first"
python -m brain_engine remediate "Surplus Review" # reviewer corrections -> apply-ready prompt
python -m brain_engine.loop --demo # INTEGRITY LOOP: re-derive tampered entries verbatim
python -m pytest -q # 7,011 tests
```
## Loop verdicts (`brain_engine/loop.py`)
`CLEAN` (0) / `REMEDIATED` (0) / `ESCALATED` (3 — the engine's refusal code).
The signature rule: tampered text/speakers are re-derived BYTE-FOR-BYTE from `Corpus.utterance_text`; **orphaned citations (no utterance behind them) are quarantined, never repaired** — no source, no guess.
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!