moss-image-prompt-enhancer
Pass
Audited by Gen Agent Trust Hub on Oct 24, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructions direct the agent to run a local Python script (
compile_prompt.py) to process and validate image layout drafts. This is a functional requirement for the skill's purpose and uses standard, non-shell command execution.- [DYNAMIC_EXECUTION]: The included test suite (test_compile_prompt.py) utilizes standard Python dynamic module loading (importlib) to execute unit tests against the compiler script within a controlled testing environment.- [INDIRECT_PROMPT_INJECTION]: The skill has a data processing surface where it transforms user-supplied image ideas into structured JSON layouts. - Ingestion points: The agent ingests user image requests, and the
compile_prompt.pyscript reads the resulting JSON draft file. - Boundary markers: The skill provides clear schema documentation in
references/layout.mdandSKILL.mdto guide the agent in generating structured data. - Capability inventory: The skill performs file read/write operations and executes the local Python script for data transformation.
- Sanitization: The compiler script implements robust sanitization, including strict schema validation, duplicate key detection, and the rejection of unsafe JSON constants like NaN and Infinity.
Audit Metadata