The Agent Skills Directory

[SAFE]: No malicious behavior or suspicious instructions were detected. The skill's behavior is consistent with its stated purpose of facilitating product planning and market analysis.
[PROMPT_INJECTION]: The skill contains an indirect prompt injection surface due to its extensive use of web search tools to research competitors and market trends. 1. Ingestion points: Web search results from sources including Google, OpenAI blogs, Reddit, and Product Hunt (SKILL.md). 2. Boundary markers: No explicit delimiters or boundary markers for external data are specified. 3. Capability inventory: The skill can write to the local file system (docs/product-concept.md) and invoke other tools (baoyu-slide-deck). 4. Sanitization: No evidence of sanitization for ingested web content. This surface is considered safe as it is a fundamental requirement of the skill's research functionality.
[COMMAND_EXECUTION]: The skill utilizes slash command syntax (e.g., /baoyu-slide-deck) to delegate tasks to other skills. This represents standard platform functionality for skill-to-skill interaction and does not involve arbitrary or malicious command execution.

product-ideation