Private Secrets

Securely store, list, and retrieve your private information like API keys, passwords, and tokens in a local JSON file.

This skill is functionally coherent but has real safety issues you should consider before installing: it saves secrets in plaintext under /workspace (not safe for sensitive keys), the package metadata does not declare that 'node' is required, and the implementation uses node -e with direct string interpolation—this allows a malicious or accidental secret name/value to execute arbitrary JavaScript on your machine. If you plan to use it: (1) only run it in a trusted, isolated environment; (2) verify Node is the correct version and present; (3) do not store high-value secrets there—use an encrypted secret manager instead; (4) inspect and/or modify the script to remove node -e usage (use a small Node script file that reads process.argv and safely serializes inputs, or use jq to edit JSON), enforce allowed characters for names, and set strict filesystem permissions on secrets.json. If you are not comfortable auditing or fixing these issues, prefer a vetted secret-management tool rather than this skill.

Type: OpenClaw Skill Name: private-secrets-1-0-0 Version: 1.0.0 The skill provides a basic secret management utility but contains critical command injection vulnerabilities in private-secrets.sh. The script passes unsanitized shell variables ($NAME and $VALUE) directly into a 'node -e' execution string, which allows for arbitrary code execution if the agent or a user provides crafted input. While the tool explicitly warns in SKILL.md that secrets are stored unencrypted in secrets.json, the lack of input sanitization makes the implementation highly risky, though there is no clear evidence of intentional malicious exfiltration.