The Agent Skills Directory

[PROMPT_INJECTION]: The skill creates a surface for indirect prompt injection through its instructions to read and process external data sources.
Ingestion points: File content ingested via Read and Grep tools, and PDF content processing (SKILL.md).
Boundary markers: The instructions do not specify the use of delimiters or 'ignore embedded instructions' warnings when reading external content.
Capability inventory: The skill environment includes file modification capabilities (Edit, Write tools) and interaction with external tools/models (Qwen MCP).
Sanitization: There is no mention of sanitizing, escaping, or validating the content retrieved from external files before it is processed by the agent.

token-conservation