Repomix Explorer

Use this skill when the user wants to analyze or explore a codebase (remote repository or local repository) using Repomix. Triggers on: 'analyze this repo',...

This skill appears to do what it says: run Repomix to pack a repo and analyze the output. Before installing/using it, consider: (1) npx repomix@latest will fetch and execute code from the npm registry — if you want stronger guarantees, pin a specific repomix version (e.g., [email protected]) or install/inspect the CLI yourself; (2) the agent will read repository files and any generated output in /tmp or the working directory — do not run it on repositories containing secrets you don't want to expose; (3) run the skill in a sandbox/container if you are worried about executing third-party code; (4) verify the repomix package source (npm package page, GitHub repo, checksum) if you need higher assurance. If you control the environment, manually installing and vetting repomix before allowing the agent to run it reduces risk.

Type: OpenClaw Skill Name: repomix-explorer Version: 0.1.0 The skill instructs the AI agent to execute shell commands (`npx repomix@latest`, `grep`) with parameters directly derived from user input (e.g., repository names, local paths, output paths, include/ignore patterns). The `SKILL.md` instructions do not include explicit directives for input sanitization, which creates a significant shell injection vulnerability. A malicious user could craft inputs containing shell metacharacters to execute arbitrary commands or write files to unintended locations. While the skill's stated purpose is benign and there is no evidence of intentional malicious behavior (such as data exfiltration or persistence), the lack of input sanitization for shell commands makes it suspicious.