Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
Claude Code Runner
v0.1.0Execute programming tasks via Claude Code using PTY-based invocation. Handles non-TTY environments, auto-responds to prompts, and manages file synchronization.
⭐ 1· 1.1k·10 current·10 all-time
byholenlin@lhl09120
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
high confidencePurpose & Capability
Name/description match the included script: PTY execution, auto-respond, file sync, and user switching. Minor metadata mismatch: registry metadata lists no required binaries, but documentation and clawhub.json require the 'claude' tool to be installed and in PATH.
Instruction Scope
Runtime instructions and the script operate on arbitrary project directories, copy them to a temp dir, change ownership, run a shell command that includes the user-supplied prompt, and sync modifications back — all expected for this use case but risky. Critically, the command is built into a single shell -c string with the raw prompt interpolated (f'... claude --print "{prompt}" ...'), enabling shell injection if the prompt or path contains special characters. The skill also requires root/sudo to chown and switch users, so a malicious or malformed prompt could execute arbitrary commands as the target user.
Install Mechanism
Instruction-only skill (no install spec). SKILL.md instructs cloning a GitHub repo and marking the script executable — standard for an instruction-only package. No downloads from unknown hosts or archives are present in the registry metadata.
Credentials
No required environment variables or credentials are declared. Optional env vars (CLAUDE_CODE_USER, CLAUDE_CODE_TIMEOUT) are documented but not required. The need for root/sudo to change ownership and perform user switching is legitimate given the feature set, but requesting that privilege increases risk and should be minimized.
Persistence & Privilege
Skill does not request persistent installation or always:true. It needs elevated privileges at runtime (root/sudo) to perform chown and su operations, which is a normal requirement for user-switching but raises risk if run on sensitive systems.
What to consider before installing
This skill is coherent with its stated purpose (running Claude in a PTY and syncing changes), but it contains a high-risk programming error: the script builds a shell command by interpolating the user-provided prompt directly into su -c "...". That makes it possible for specially crafted prompts or paths to execute arbitrary shell commands as the target user. Before installing or running this skill: (1) review the scripts locally — do not run it on production hosts; (2) run it in an isolated container or VM with no sensitive data and minimal privileges; (3) if you plan to use it, patch the script to avoid shell interpolation (use subprocess with an argument list, avoid shell=True/su -c with unescaped data, safely escape paths/prompts); (4) avoid running it as root when possible — run with least privilege and a dedicated user; (5) verify the origin of the repository (the SKILL.md recommends cloning from GitHub but the registry source is unknown). Because of the shell-injection risk and the file-write privileges, treat this skill as suspicious until the unsafe command construction is fixed.Like a lobster shell, security has layers — review code before you run it.
latestvk97fye1vmayadnq7cbthvhbmk981zme2
License
MIT-0
Free to use, modify, and redistribute. No attribution required.