The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests data from a local file (~/.claude/company-context.md) which acts as an attack surface for indirect prompt injection. The skill includes a comprehensive anonymization-protocol.md reference to sanitize this data before external tool use. Ingestion points: SKILL.md. Boundary markers: Internal status notes. Capability inventory: No dangerous local commands or network scripts present. Sanitization: Detailed anonymization rules for financial and personal data.

context-engine