The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted external content such as customer interview transcripts, research papers, and project source code for analysis. While no immediate exploit was found, this creates a surface for indirect prompt injection. Ingestion points are located in product-manager-toolkit/SKILL.md, research-summarizer/SKILL.md, and code-to-prd/SKILL.md. The skill possesses capabilities for file system modification and script execution.
[DYNAMIC_EXECUTION]: Modules like spec-to-repo and saas-scaffolder generate functional source code and directory structures from natural language requirements. This is the intended purpose of these tools and relies on established templates.

product-skills