The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted input from task descriptions and external files, which represents an indirect prompt injection surface.
Ingestion points: Task description argument, AGENTS.md, design documents in design/gdd/, and historical sprint data in production/sprints/.
Boundary markers: No delimiters are used to separate ingested data from agent instructions.
Capability inventory: The skill is restricted to Read, Glob, and Grep tools (read-only access).
Sanitization: No content validation or sanitization is performed on the ingested data.
[SAFE]: No malicious patterns such as data exfiltration, hardcoded credentials, or remote code execution were detected. The skill's functionality is limited to information gathering and report generation within the local environment.

estimate