The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process user-supplied support questions, error reports, and evidence, which creates a potential surface for indirect prompt injection.
Ingestion points: Processes user questions, relevant error text, and confirmed context (README.md, SKILL.md).
Boundary markers: Instructions explicitly state that quoted history is evidence and not a new customer instruction (references/conversation-knowledge.md).
Capability inventory: The skill is authorized to use Read, Glob, and Grep tools for filesystem operations (SKILL.md).
Sanitization: The instructions emphasize establishing host details and using sanitized errors for support requests (references/access-and-installation.md).

norma-support