Generate and execute safe shell commands to explore codebases in parallel.
npx skills add https://github.com/Graffioh/dotfiles --skill codebase-ripper
Please help me install this Agent Skill. Skill: codebase-ripper Source: https://github.com/Graffioh/dotfiles/tree/main/pi/agent/skills/codebase-ripper Command: npx skills add https://github.com/Graffioh/dotfiles --skill codebase-ripper
This Skill enables rapid, iterative exploration of a codebase by generating and executing a large suite of safe shell commands to build a comprehensive context, reducing manual inspection time.
Run the codebase ripper against a project with a task such as "Map the authentication flow" using the tool's run script: ~/.pi/agent/skills/codebase-ripper/run.sh "Map the authentication flow" -d /path/to/project
High-intent search queries and answers about installing and using this skill.
Automating codebase exploration involves generating and executing a broad set of safe shell commands to collect contextual signals. It maps project structure and dependencies across iterative passes, reducing manual inspection time when onboarding to unfamiliar repositories.
Automated codebase exploration works across languages by enforcing a strict allowlist of safe commands. It runs vetted tasks in parallel to extract relevant context, making it ideal for auditing patterns before implementing features or tests in diverse repositories.
You can limit shell command execution during codebase exploration by configuring max-iterations and token-budget parameters. These settings control iterative passes, ensuring the extraction of structured context remains within your defined processing limits.
The best way to gather pre-coding context safely is by using an automated tool that enforces a strict allowlist of shell commands. This ensures only vetted, safe commands run in parallel to extract context without risking repository integrity.
To run parallel commands for context extraction, execute the tool's run script with your target project directory path. The script generates a broad command set, executing vetted tasks in parallel to map the codebase structure efficiently.
jbovetSearch, list, and inspect files in read-only mode across filesystems.
jctannerExplore checkouts directories to generate component-map.json from breadcrumb signals.
VMAFxRun multiple static analysis tools and merge findings into one report.
catlog22Orchestrate parallel codebase explorations into plan.json and TASK-*.json files.
