The Agent Skills Directory

[PROMPT_INJECTION]: The skill presents an indirect prompt injection surface, as it is designed to ingest and process information from untrusted external sources (web search results and third-party APIs).
Ingestion points: Data entering the agent context from WebSearch, WebFetch, Perplexity, and Gemini agents as described in SKILL.md.
Boundary markers: No delimiters or specific instructions to ignore embedded commands are present in this scaffold.
Capability inventory: Parallel execution of sub-tasks via the Task tool and web search capabilities.
Sanitization: No evidence of escaping or validation of external content is defined in the current workflow.
[SAFE]: The skill references environment variables (PERPLEXITY_API_KEY, GOOGLE_API_KEY) for standard API authentication required for its research functions.
[SAFE]: External links to GitHub are provided for documentation and design patterns, which are informative and do not involve automated code execution.

research-parallel