image-generation
Pass
Audited by Gen Agent Trust Hub on Nov 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's primary function is to process user-generated prompts to produce images, which creates an inherent attack surface where malicious input could attempt to influence the agent's behavior or the underlying model's output.
- Ingestion points: User-defined text inputs for subject, style, lighting, and composition specified in
SKILL.md(Step 2) andSKILL.toon(S[5]). - Boundary markers: No specific delimiting or instruction-ignoring markers are documented; user input is shown being passed directly as a string to the
ask-geminicommand. - Capability inventory: The skill utilizes CLI tools (
ask-gemini,claude mcp) and manages local files (saving images and JSON metadata) as defined inSKILL.md(Step 1 and Step 5). - Sanitization: There are no instructions or automated checks for sanitizing or escaping the user-provided prompt content before execution.
- [COMMAND_EXECUTION]: The skill requires the use of local command-line tools to interact with MCP servers and generation models.
SKILL.md(Step 1 and Step 4) instructs users to executeclaude mcp listto check server status andask-geminito run the generation requests.
Audit Metadata