Skill Factory

Create, evaluate, improve, benchmark, and publish OpenClaw skills. Use when building a new skill from scratch, iterating on an existing skill, running evals...

This skill appears to do what it says: scaffold skills, validate frontmatter, scan installed skill SKILL.md files to synthesize patterns, and package skill folders. Before running: 1) Inspect the bundled scripts (they are included) to confirm behavior (they are plain Python). 2) Be aware analyze_patterns.py will read SKILL.md and related files under your skills directories (defaults include ~/.openclaw/workspace/skills and an openclaw path under ~/.nvm); avoid running it against directories that contain secrets you don't want aggregated. 3) When using init_skill or package_skill, pass explicit paths rather than relying on example find commands so you control what gets created or packaged. 4) If you plan to run these tools in an automated or shared environment, run them in a sandbox first to verify outputs. Overall the package is internally consistent and does not ask for extra credentials or perform unexpected network calls.

Type: OpenClaw Skill Name: skill-factory Version: 1.0.0 The skill bundle is classified as suspicious due to inherent risks associated with its core functionality of creating, packaging, and analyzing other skills, which involves executing shell commands and broad file system access. Specifically, the `SKILL.md` instructs the AI agent to execute Python scripts with user-provided input (e.g., `<skill-name>`), creating a potential prompt injection vector against the agent, even though the `init_skill.py` script itself includes input sanitization. Additionally, the `SKILL.md` uses a `find` command to locate its own scripts, which could be vulnerable to path manipulation if a malicious script with the same name were planted in an earlier search path. While these capabilities are necessary for the skill's stated purpose, they represent significant vulnerabilities that could be exploited, though there is no clear evidence of intentional malicious behavior like data exfiltration or backdoors within the provided code.