tech-seo-audit
Fail
Audited by Gen Agent Trust Hub on May 20, 2026
Risk Level: HIGHCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a local Python script using the command
python "scripts/tech-seo-auditor.py" --url {url}. Because the{url}argument is provided directly by the user without sanitization, an attacker can use shell metacharacters to perform command injection and execute arbitrary code on the host system. - [DATA_EXFILTRATION]: The skill is designed to read multiple sensitive files from the user's home directory, specifically under
~/.claude-marketing/brands/, includingprofile.json,_manifest.json, and SOPs. While this is the intended functionality, the combination of local file access and command execution capabilities creates a high risk for data exfiltration. - [INDIRECT_PROMPT_INJECTION]: The skill analyzes external website content via a user-provided URL. This creates a surface for indirect prompt injection where a malicious website could embed instructions in HTML, headers, or metadata to influence the agent's audit report or subsequent actions.
- Ingestion points: User-provided URL processed in SKILL.md.
- Boundary markers: None identified.
- Capability inventory: Subprocess execution of
scripts/tech-seo-auditor.pyand read access to~/.claude-marketing/filesystem. - Sanitization: No input validation or content sanitization is described.
Recommendations
- AI detected serious security threats
Audit Metadata