audience-research
Pass
Audited by Gen Agent Trust Hub on Nov 17, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a local Python script
brand_context.pyusingpython3to initialize brand context. The script is accessed via a relative path from the skill directory. - [DYNAMIC_EXECUTION]: The skill uses a
general-purposetool to programmatically spawn and manage multiple subagents in parallel to handle distinct data mining tasks across different sources. - [EXTERNAL_DOWNLOADS]: The skill fetches data from multiple external platforms, including Reddit, Amazon, YouTube, and specialized review sites (G2, Capterra). It uses tools such as Firecrawl, Exa, and ScrapeCreators to perform these retrievals as part of its research functionality.
- [INDIRECT_PROMPT_INJECTION]: The skill's core function is to ingest and process untrusted third-party content from social media and review platforms. This content is captured verbatim, creating a potential vector for indirect prompt injection where malicious instructions embedded in reviews could influence the agent's behavior during synthesis.
- Ingestion points: External web data retrieved from various social media and review platforms during Phase 2.
- Boundary markers: The skill structures data into a YAML-formatted "evidence pack" and provides a rubric for signal vs. noise in
voc-mining-judgment.mdto help filter low-quality data. - Capability inventory: The skill possesses capabilities for file system writes (
audience.md,voc.md), local command execution (python3), and subagent creation. - Sanitization: The skill specifically mandates "verbatim" capture of text (preserving typos and slang), which means no sanitization is applied to the raw external input before it is processed by the model for synthesis.
Audit Metadata