seo-image-gen
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs the user or agent to install an external extension via a local shell script (
./extensions/banana/install.sh) and relies on an external Model Context Protocol (MCP) server documentation pointing to a public GitHub repository. - [COMMAND_EXECUTION]: The skill frequently invokes local Python scripts and command-line utilities to perform its core functions:
- Executes internal scripts for preset management (
presets.py), cost tracking (cost_tracker.py), and fallback image generation (generate.py). - Utilizes system-level tools such as
magick(ImageMagick) andffmpegfor image conversion, resizing, and manipulation. - [INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection due to its processing of untrusted external data.
- Ingestion points: Reads site metadata from the
.seo-cache/directory and accepts arbitrary text descriptions from user commands to build image generation prompts. - Boundary markers: The prompt construction logic does not explicitly define delimiters or specific instructions to ignore malicious directives embedded within the ingested metadata or user descriptions.
- Capability inventory: The skill has access to shell execution for image processing and local script execution.
- Sanitization: No formal input validation or sanitization mechanisms are described for the data processed from the cache or user inputs.
Audit Metadata