precision edit
editingworkflowcopywebsite
/install muscle precision-edit Run in a Soma session to add to your project.
curl -sO https://raw.githubusercontent.com/meetsoma/community/main/muscles/precision-edit.md Download and copy to .soma/muscles/precision-edit.md — or view on GitHub ↗
Details
Precision Edit
TL;DR
Precision Edit — before editing any multi-line file, extract content with line numbers first. (1) Run relevant verify script (soma-verify.sh copy, soma-verify-styles.sh) to find issues. (2) grep -n the target patterns to get exact line numbers. (3) read --offset --limit the exact ranges to see surrounding whitespace. (4) Plan all edits as a line-numbered checklist before touching anything. (5) Edit with exact whitespace matches. Never edit blind — the match must be verified by read first.
The Pattern
1. SCAN — run verify script to find what's wrong
2. LOCATE — grep -n to get line numbers for each issue
3. READ — read --offset --limit to see exact content + whitespace
4. PLAN — list every edit: file:line, old text, new text
5. EDIT — execute edits with exact whitespace from step 3
6. VERIFY — re-run verify script to confirm fixesWhy
edit requires exact text matching including whitespace. Guessing whitespace from memory or grep output fails. The only reliable path:
grep -nfinds the line numberread --offset --limitshows the exact content with all whitespace- Copy the exact text from read output into the edit
Anti-patterns
- ❌ Editing based on line numbers from a previous read (content shifts after edits)
- ❌ Guessing indentation from grep output (grep strips context)
- ❌ Multiple edits without re-reading between them (line numbers drift)
- ❌ Editing without running the verify script first (might fix the wrong thing)
Related
soma-verify.sh copy— finds stale copy, counts, framing issuessoma-verify-styles.sh— finds raw CSS values, inconsistent typographysoma-verify.sh website— docs sync verification