/fix
Watch a bug video, diagnose the issue, fix the code, and raise a PR
$ npx -y skills add mnvsk97/eyeroll --agent claude-codeHow it fires
How this command gets triggered: by you, by Claude, or both.
- Fires itselfClaude auto-loads it when your prompt matches the work.
- You can call itInvoke it directly when you want it.
- Slash command
/fix
Context preview
What this command does when you run it.
Watch a bug video, diagnose the issue, fix the code, and raise a PR
Command definition
fix.mddescription: Watch a bug video, diagnose the issue, fix the code, and raise a PR
argument-hint: '<url-or-path> [--context "..."]'
allowed-tools: Bash, Read, Write, Edit, Glob, Grep
Watch a video showing a bug, diagnose it, fix the code, and raise a PR.
Raw arguments: `$ARGUMENTS`
Step 1: Analyze the video
Codebase context is auto-discovered from CLAUDE.md, AGENTS.md, .cursorrules, etc.
eyeroll watch $ARGUMENTS --verbose
If the user mentioned context in conversation, add `-c "<their words>"`.
Step 2: Triage
Read the report. Check the **Metadata** block first:
- If **scope is out-of-context**, warn the user — the video doesn't relate to this codebase
- If **actionable is no**, tell the user — there's nothing concrete to fix here
- If **category is not bug**, note this — the fix command is designed for bugs
Check **severity** to set urgency expectations (critical/moderate/low).
Step 3: Diagnose
Read the report. Extract:
- **Reproduction steps** (from the dedicated section)
- Error messages (exact text)
- URLs and routes visible
- Fix directions (search patterns, hypotheses)
Search the codebase using the suggested patterns with `Grep` and `Glob` tools. Cross-reference with the video observations.
If context mentions a regression ("after PR #X", "since last deploy"):
gh pr view <number>
git log --since="1 week ago" -- <identified-files>
Step 4: Fix
Implement the fix. Keep changes minimal and focused.
Step 5: Test
Run the project's test suite. If a test would have caught this bug, write one.
Step 6: PR
Create a branch, commit, push, and open a PR:
gh pr create --title "Fix: <brief description>" --body "$(cat <<'EOF'
## Summary
<what was broken and why>
## Bug Evidence
- Video: <source URL if available>
- Error: <exact error message>
## Changes
- <file>: <what changed>
## Test Plan
- <how to verify>
EOF
)"
If diagnosis confidence is low, present findings and ask the user before implementing. Create the PR as a draft if confidence is medium or lower.
Read more
description: Watch a bug video, diagnose the issue, fix the code, and raise a PR argument-hint: '<url-or-path> [--context "..."]' allowed-tools: Bash, Read, Write, Edit, Glob, Grep
Watch a video showing a bug, diagnose it, fix the code, and raise a PR.
Raw arguments: `$ARGUMENTS`
Step 1: Analyze the video
Codebase context is auto-discovered from CLAUDE.md, AGENTS.md, .cursorrules, etc.
eyeroll watch $ARGUMENTS --verbose
If the user mentioned context in conversation, add `-c "<their words>"`.
Step 2: Triage
Read the report. Check the **Metadata** block first:
- If **scope is out-of-context**, warn the user — the video doesn't relate to this codebase
- If **actionable is no**, tell the user — there's nothing concrete to fix here
- If **category is not bug**, note this — the fix command is designed for bugs
Check **severity** to set urgency expectations (critical/moderate/low).
Step 3: Diagnose
Read the report. Extract:
- **Reproduction steps** (from the dedicated section)
- Error messages (exact text)
- URLs and routes visible
- Fix directions (search patterns, hypotheses)
Search the codebase using the suggested patterns with `Grep` and `Glob` tools. Cross-reference with the video observations.
If context mentions a regression ("after PR #X", "since last deploy"):
gh pr view <number> git log --since="1 week ago" -- <identified-files>
Step 4: Fix
Implement the fix. Keep changes minimal and focused.
Step 5: Test
Run the project's test suite. If a test would have caught this bug, write one.
Step 6: PR
Create a branch, commit, push, and open a PR:
gh pr create --title "Fix: <brief description>" --body "$(cat <<'EOF' ## Summary <what was broken and why> ## Bug Evidence - Video: <source URL if available> - Error: <exact error message> ## Changes - <file>: <what changed> ## Test Plan - <how to verify> EOF )"
If diagnosis confidence is low, present findings and ask the user before implementing. Create the PR as a draft if confidence is medium or lower.
AI eyes that roll through video footage — watch, understand, act. eyeroll is a Claude Code plugin that analyzes screen recordings, Loom videos, YouTube links, and screenshots, then helps coding agents fix bugs, build features, and create skills.
Other commands on eyeroll.
- /api-watch
Analyze a video URL or local file using the eyeroll hosted API.
Open command - /history
List past video analyses from the cache
Open command - /init
Set up eyeroll for this project — pick a backend, install dependencies, and generate codebase context
Open command - /watch
Analyze a video, screen recording, or screenshot and produce a structured report
Open command

