Skills
Navigate
skills/reason-machines/hermes-skills/openclaw-tool-call-viewer-2026/Gen Agent Trust Hub

openclaw-tool-call-viewer-2026

Fail

Audited by Gen Agent Trust Hub on Sep 12, 2026

Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill provides instructions to clone a software repository from an unverified external source: https://github.com/reedchris43/openclaw-tool-call-viewer-2026.git.
  • [REMOTE_CODE_EXECUTION]: The skill facilitates the execution of code downloaded from an unverified repository by providing commands to serve it locally using Python, Node.js, or PHP, allowing third-party code to run in the user's local environment.
  • [COMMAND_EXECUTION]: The skill includes shell commands (git clone, python3 -m http.server, npx http-server, php -S) that perform network operations and interact with the local file system.
  • [INDIRECT_PROMPT_INJECTION]: The viewer implementation is vulnerable to Cross-Site Scripting (XSS) because it processes untrusted external data without sanitization.
  • Ingestion points: External JSON session data is loaded via the loadSession function from local files or remote API endpoints.
  • Boundary markers: None present; the code does not distinguish between data and executable instructions within the JSON properties.
  • Capability inventory: The skill uses fetch for network data retrieval and innerHTML for DOM manipulation.
  • Sanitization: Absent; the renderToolCalls function directly interpolates object properties like call.tool_name and call.id into HTML strings, allowing an attacker to execute malicious scripts if they can control the session data content.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Sep 12, 2026, 06:50 PM
Security Audit — agent-trust-hub — openclaw-tool-call-viewer-2026