The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes conversation context and codebase understanding as input for PRD generation, which constitutes a surface for indirect prompt injection. 1. Ingestion points: Conversation context and codebase understanding (referenced in SKILL.md). 2. Boundary markers: None identified; input data is not wrapped in specific delimiters. 3. Capability inventory: The agent is instructed to publish output to the project's issue tracker. 4. Sanitization: None identified. This surface is considered safe as it is inherent to the skill's primary utility and the output is constrained by a fixed markdown template.
[SAFE]: No malicious obfuscation, remote code execution patterns, or unauthorized credential access attempts were detected. The reference to a setup command for environment configuration follows standard modular design practices.

to-prd