openclaw-marketing-skills
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from external URLs (e.g., landing page audits via 'Analyze https://example.com/demo'), which presents a surface for indirect prompt injection attacks where malicious instructions on those sites could influence the agent's behavior.
- Ingestion points: User-provided URLs for conversion and SEO audits in SKILL.md.
- Boundary markers: None identified; the skill does not appear to use delimiters to isolate external content.
- Capability inventory: The skill has the ability to write JSON data files to the '.agents/' directory and perform network API calls to Google Ads, Meta Ads, and Search Console.
- Sanitization: No explicit sanitization or filtering of the fetched website content is mentioned.
- [EXTERNAL_DOWNLOADS]: The skill instructions direct users to install components from an external GitHub repository (LeoYeAI/openclaw-marketing-skills), which introduces a dependency on external sources.
- [COMMAND_EXECUTION]: The troubleshooting and setup sections involve executing shell commands to verify installation, check file permissions, and run diagnostic Python code snippets.
Audit Metadata