The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection attacks.
Ingestion points: The skill uses the WebFetch tool to retrieve content from arbitrary external websites (SKILL.md).
Boundary markers: There are no instructions to use delimiters or ignore embedded prompts within the fetched data before synthesis.
Capability inventory: The skill has Write tool access, which it uses to save results to the local workspace/ folder (SKILL.md).
Sanitization: The skill lacks explicit sanitization or validation logic for data retrieved from external sources.

quick-research