seo-image-gen
Fail
Audited by Gen Agent Trust Hub on Oct 28, 2026
Risk Level: HIGHCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill invokes an external installation script (
./extensions/banana/install.sh) and several Python scripts (presets.py,cost_tracker.py,generate.py) via theBashtool. - [REMOTE_CODE_EXECUTION]: The skill relies on unbundled code sourced from a personal GitHub repository (
AgriciDaniel/claude-seo). Executing these non-vetted scripts provides a direct path for arbitrary code execution on the host machine. - [EXTERNAL_DOWNLOADS]: The skill instructs the user to download prerequisites and reference materials from an untrusted external GitHub repository.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user input from the
<description>argument. - Ingestion points: The
<description>argument provided by the user in the command line. - Boundary markers: No specific boundary markers or instructions to ignore embedded commands are present in the Reasoning Brief construction logic.
- Capability inventory: Access to
Bash(script execution),Write(file creation), andWebFetch(network requests). - Sanitization: No evidence of sanitization or filtering of the user-provided description before it is interpolated into the generation prompt.
Recommendations
- AI detected serious security threats
Audit Metadata