# Conformance Loop

> Iteratively improve Fallow analysis accuracy by comparing it with competing tools and verified source truth across a stable real-world corpus.

- Skill: `fallow-rs/conformance-loop` (Agent Skill)
- Install (CLI): `npx skillmds@latest add fallow-rs/conformance-loop`
- Raw SKILL.md: https://api.skillmd.com/api/skills/fallow-rs/conformance-loop/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: fallow-rs (https://skillmd.com/u/fallow-rs)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/fallow-rs/conformance-loop

---


# Conformance loop

1. Define the capability and stable project corpus.
2. Run Fallow and comparison tools with documented equivalent settings.
3. Manually verify disagreements against source.
4. Classify true positive, false positive, false negative, or model difference.
5. Record the classification in `tests/conformance/adjudications.json` before
   writing the fix. A verdict that lives only in a session or in `.plans/` is
   lost, and the daily lane then re-reports the same disagreement as
   unadjudicated forever.
6. Implement one general correction with a regression fixture.
7. Re-run the full corpus and retain only net improvements.
8. Run `review`.

Competitor output is a lead, not ground truth.