The Agent Skills Directory

[SAFE]: The skill performs legitimate document processing by reading local Markdown or PDF files and writing structured JSON output to the project's directory. All filesystem operations are consistent with the skill's stated purpose of preparing data for a presentation tool.
[COMMAND_EXECUTION]: The skill includes instructions to suggest a shell command for the user to execute a local Python script, nano_banana.py. This script is part of the local environment and is used to generate slide images from the produced specifications. This behavior is functional and transparent to the user.
[PROMPT_INJECTION]: The skill processes untrusted content from user-provided reports, which constitutes an indirect prompt injection surface. This is a standard characteristic of document-analysis skills and does not represent an active exploit.
Ingestion points: Report content is loaded into the agent's context via the report_path argument using the Read tool.
Boundary markers: There are no explicit delimiters or instructions provided to isolate the report content from the system prompts.
Capability inventory: The skill has access to local filesystem read/write tools and can suggest local shell command execution.
Sanitization: No specific sanitization or filtering logic is defined for the content extracted from the source documents.

deck-prompt