The Agent Skills Directory

[PROMPT_INJECTION]: The skill contains instructions that ingest untrusted data from the local project environment to provide contextual recommendations, establishing a surface for indirect prompt injection.
Ingestion points: SKILL.md (Initial Assessment) instructs the agent to read .agents/product-marketing.md, .claude/product-marketing.md, or product-marketing-context.md.
Boundary markers: Absent. The instructions do not define delimiters or specify that the content of these files should be treated as untrusted data.
Capability inventory: The skill defines capabilities for generating marketing copy, recommending triggers, and designing user flows based on the provided context (SKILL.md).
Sanitization: Absent. There are no instructions to sanitize, validate, or filter the content of the external marketing context files before the agent processes them.

popups