The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to run local Python scripts such as merge_pdfs.py, build_html.py, setup_notion.py, and generate_images.py to automate product assembly. It also references system-level tools like pandoc and xelatex for document generation.
[DATA_EXPOSURE]: The skill exposes the local system's directory structure by using hardcoded absolute file paths (E:\Dev2\DigitalProduct\) to retrieve reference materials and create project scaffolds.
[INDIRECT_PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface due to its multi-step architecture where one phase generates instructions (BUILD_PROMPT.md) to be executed by the agent in a later session. 1. Ingestion points: Untrusted user input enters the workflow via product discovery (Phase 1) and content production (Phase 4). 2. Boundary markers: The skill does not define specific delimiters or instructions to prevent the agent from obeying instructions embedded in the user-supplied content. 3. Capability inventory: The skill has access to local script execution and file system write operations across the directory structure. 4. Sanitization: There is no evidence of sanitization or validation of external content before it is interpolated into build instructions or final deliverables.

kai-product-maker