The Agent Skills Directory

[SAFE]: The skill performs legitimate document transformation tasks by reading local project files and writing a structured markdown output. The operations are consistent with the stated purpose of a Product Manager AI agent.
[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface because it processes untrusted external data (product plans and market research documents) and interpolates this content into its output. However, the potential impact is negligible as the skill lacks high-risk capabilities like network access or arbitrary command execution. Evidence includes:
Ingestion points: Reads content from product-plan-*.md, competitive-analysis.md, and market-research.md using the Read tool.
Boundary markers: The instructions do not specify the use of XML tags or explicit 'ignore instructions' delimiters for the interpolated content.
Capability inventory: Limited to directory creation (mkdir -p docs) and file writing (Write tool) within the local environment.
Sanitization: No explicit validation or sanitization steps are defined for the imported text.

prd-generator