What is the primary benefit of the Observe Before Editing skill?

It prevents the AI from making guesses about why code failed by forcing it to check logs, cache files, and manual error outputs before proposing any code changes.

Is this skill only for the Maestro framework?

While it contains specific patterns for Maestro, the 'observe before editing' philosophy is a best practice that improves Claude's debugging accuracy across any development task.

Which specific paths does this skill monitor?

The skill specifically instructs Claude to check .maestro/cache/, .maestro/cache/learnings/, and to distinguish between local and global (~/.maestro/) directories.

How does this help with silent errors?

By requiring manual execution of failing commands and log tailing, it uncovers errors like token limit breaches that may not be immediately visible in the standard execution flow.

Observe Before Editing

Name: Observe Before Editing
Author: scooter-lacroix

byscooter-lacroix

0•

分析と監視

Enforces an observation-first debugging workflow by verifying system outputs and logs before making code changes.

This skill implements a rigorous 'verify-before-act' pattern for Claude Code, specifically optimized for the Maestro framework. It prevents premature and incorrect code edits by requiring the model to first validate the system's actual state—checking for the existence of cache files, inspecting logs, and manually reproducing errors. By focusing on empirical evidence from directories like .maestro/cache/ and differentiating between local and global paths, it ensures that bug fixes are targeted and based on real-world behavior rather than assumptions.

主な機能

01Manual error reproduction steps to reveal invisible failure modes

02Automated verification of .maestro/cache/ directories and learning files

03Artifact Index verification to confirm successful indexing operations

04Validation of local vs. global project paths to prevent path confusion

05Mandatory log inspection using tail and cat before editing

060 GitHub stars

ユースケース

01Investigating missing learning artifacts or indexing errors

02Identifying invisible token limit errors hidden in system logs

03Debugging silent failures in Maestro hooks or automation scripts

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add scooter-lacroix/maestro observe-before-editing

For use in Claude.ai and ChatGPT

Download Skill