The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external content such as source code, tests, and Git history, which can serve as a vector for indirect prompt injection.
Ingestion points: The agent is instructed in SKILL.md to gather full context by reading code, surrounding modules, tests, comments, and git history using tools like Read, Grep, and Glob.
Boundary markers: The instructions lack specific guidance on using delimiters or "ignore embedded instructions" warnings when handling content from external files.
Capability inventory: The skill is configured with access to Bash, Read, Grep, and Glob, providing a significant capability set if an injection were successful.
Sanitization: There are no requirements defined for sanitizing, escaping, or validating the content retrieved from the file system before the agent processes it for explanation.

explain