agentleFS
sophonfinance-wqsophonfinance-wq/finance-automation-portfolio/.claude/skills/triangulate/SKILL.md

triangulate

Skill16 starsChanged 44 days agoMITView on GitHub

Run Triangulate (multi-agent AI validation: preparer/reviewer/specialist/deterministic auditor/human gate), the adversarial hallucination demo, or the review loop where arithmetic self-heals and judgment escalates.

  • Reads credentials

3 other projects use a version of this.

Adopted by

Most recently taken up by

The file

Files
SKILL.md

Triangulate — AI Validation Framework

Package triangulate in ai-validation-framework/. Offline mock reviewer by default; live Claude reviewer (AnthropicReviewer, stdlib urllib) needs ANTHROPIC_API_KEY.

Commands (run from ai-validation-framework/)

python -m triangulate --demo-adversarial     # inject $49k fake revenue; 6 CRITICAL tie-out breaks; FAIL exit 1
python -m triangulate --sample defective     # AI-assumption + hardcoded-cell sample
python -m triangulate.loop --sample adversarial   # REVIEW LOOP: FAIL -> PASS in 1 turn (B5 re-derived), exit 0
python -m triangulate.loop --sample defective     # arithmetic fixed, judgment ESCALATED, exit 1
python -m pytest -q                          # 8,320 tests

Loop boundary (triangulate/loop.py)

Remediates ONLY tie-out breaks on formula cells, re-derived from the cell's own formula ON A CLONE (new workpaper version per turn — never violates the hash-enforced read-only guard). AI assumptions / hardcoded cells / missing cells stay in the human fix packet. "The loop clears arithmetic; it cannot manufacture authority."

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.