The Agent Skills Directory

[NO_CODE]: The skill is composed entirely of Markdown documentation, providing frameworks and templates for business diagnostics. It does not include any scripts, binary executables, or automated command-line instructions, effectively eliminating common execution-based attack vectors.- [INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it is designed to process and analyze data from various external files and user-supplied metrics. (1) Ingestion points: products.md, brand-master.md, learnings.jsonl, metrics.md, audience.md, and offers.md. (2) Boundary markers: The skill mentions a global interaction protocol for context handling but does not define specific delimiters for untrusted file content within the instructions. (3) Capability inventory: The skill is restricted to reading local files and writing structured logs to learnings.jsonl. It lacks network access or arbitrary command execution capabilities. (4) Sanitization: No specific input sanitization or filtering logic is defined in the provided instructional files. The overall risk is low due to the analytical nature of the agent's tasks.

afa-diagnose