The Agent Skills Directory

[SAFE]: The skill defines a text-based workflow for project management. Analysis of the instructions and evaluation criteria shows no attempts at prompt injection, unauthorized data access, or obfuscation.
[PROMPT_INJECTION]: The skill involves processing potentially untrusted data from stakeholder notes and tickets, which represents an indirect prompt injection surface.
Ingestion points: Stakeholder notes, requirement documents, and support tickets mentioned in SKILL.md.
Boundary markers: Not explicitly defined in the workflow steps.
Capability inventory: The skill is limited to generating text reports and does not reference any tools capable of file system modification, network communication, or code execution.
Sanitization: No specific input sanitization or validation logic is present.
[DATA_EXPOSURE]: The skill references .agents/project-context.md for project context. This is a standard practice for project-aware agents and is limited to the local project environment.

pm-plan-requirements-brief