Skill Explorer

Systematically discover, evaluate, compare, and assess OpenClaw skills to find the right, safe, and high-quality option for your specific task.

This skill appears to do what it says: search for, download to /tmp, and statically analyze other skills. Before using it, consider the following: (1) it assumes the presence of clawhub and jq but does not declare them — install those from trusted sources or the script will fail; (2) the helper will download arbitrary skills (clawhub install), so run it in a sandbox or VM and avoid running with elevated privileges; (3) downloaded skills may include install hooks or scripts — inspect files before executing anything inside them; (4) the script suggests forcing installs for 'suspicious' skills — avoid --force unless you understand the risk; (5) if you need higher assurance, manually review any network endpoints or hardcoded tokens found by the scanner and run antivirus/static-analysis tools on downloaded artifacts.

Type: OpenClaw Skill Name: skill-explorer Version: 1.0.0 The skill-explorer bundle is designed for systematically discovering, evaluating, and performing safety assessments of other OpenClaw skills. Its `SKILL.md` provides detailed instructions for an AI agent to conduct a security review, including downloading skills to a temporary directory (`/tmp`), listing their files, and performing static analysis using `grep` for suspicious patterns (e.g., `eval`, `exec`, `https?://`). The `scripts/explore.sh` script implements these analysis steps. All actions are consistent with its stated purpose as a security analysis tool, and there is no evidence of malicious intent, data exfiltration, persistence, or prompt injection designed to subvert the agent's core function or harm the host system.