The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection attack surface as it processes untrusted data from the internet to perform its primary function.
Ingestion points: The skill uses WebSearch and WebFetch tools in Step 3 and Step 5 of the SKILL.md file to retrieve complaints and market data from Reddit and other public websites.
Boundary markers: No explicit delimiters or instructions are provided to the agent to treat external content as data only or to ignore instructions contained within that content.
Capability inventory: The skill utilizes Read and Write tools to generate and store the idea-shortlist.json file in the project root.
Sanitization: There is no evidence of sanitization, filtering, or validation of the web-fetched content before it is incorporated into the agent's brainstorming and scoring logic.

idea-generator