The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection due to its operational requirements.
Ingestion points: The skill is instructed to read existing documents from the project and use WebSearch and WebFetch to gather context for drafting policies (SKILL.md).
Boundary markers: There are no explicit instructions or delimiters used to separate the content being read/fetched from the agent's core instructions, nor are there warnings to ignore embedded commands (SKILL.md).
Capability inventory: The skill has access to high-privilege tools including Bash (shell execution), Write (file system modification), and WebFetch (network operations), as defined in its allowed tools list.
Sanitization: The instructions do not describe any mechanisms for validating, filtering, or sanitizing the data retrieved from external sources before processing.

terms-privacy