agentic-seo-skill
Fail
Audited by Gen Agent Trust Hub on Jun 16, 2026
Risk Level: CRITICALREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill instructs users to install its components by piping a remote shell script directly from an external GitHub repository into the bash interpreter (curl | bash). This execution method bypasses all verification and security reviews of the code being run.
- [COMMAND_EXECUTION]: The installation instructions for Windows suggest bypassing the PowerShell execution policy using '-ExecutionPolicy Bypass', which removes standard security protections designed to prevent the execution of untrusted scripts on a system.
- [EXTERNAL_DOWNLOADS]: The skill relies on downloading and executing multiple Python scripts and dependencies from a third-party, unverifiable GitHub repository (Bhanunamikaze/Agentic-SEO-Skill) which is not identified as a trusted vendor.
- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection as part of its core workflow.
- Ingestion points: Untrusted content enters the agent context through 'scripts/fetch_page.py' and 'scripts/crawl_audit.py' which fetch arbitrary HTML from user-provided URLs.
- Boundary markers: There are no defined delimiters or instructions to the agent to ignore embedded commands within the fetched HTML.
- Capability inventory: The skill performs file system writes (generating JSON and HTML reports), executes subprocesses (Python scripts), and makes network requests.
- Sanitization: No validation or sanitization of the external HTML content is implemented before it is processed by the LLM for reasoning.
Recommendations
- HIGH: Downloads and executes remote code from: https://raw.githubusercontent.com/Bhanunamikaze/Agentic-SEO-Skill/main/install.sh - DO NOT USE without thorough review
- AI detected serious security threats
Audit Metadata