run-organic-social-platform-research
Pass
Audited by Gen Agent Trust Hub on Nov 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from various social media platforms (X, YouTube, LinkedIn, TikTok, Instagram, Bluesky) and general web search results.
- Ingestion points: Data enters the agent's context through web search results, social media post scraping, and the ingestion of user comments from target accounts (SKILL.md, sections 1, 2, and 3).
- Boundary markers: The skill implements analysis frameworks—such as tagging comments by specific intent (pain, question, objection) and filling out structured 'How it works' cards—which help the agent contextualize external data as information rather than instructions.
- Capability inventory: The skill's actions are restricted to retrieving data via search tools and writing research results to local Markdown files; it lacks high-risk capabilities such as arbitrary command execution, privilege escalation, or access to sensitive system configurations.
- Sanitization: The instructions do not mandate technical sanitization of the raw ingested strings, relying instead on the agent's categorization logic to manage the data.
- [EXTERNAL_DOWNLOADS]: Fetches public feed data and engagement metrics from the Bluesky API (
public.api.bsky.app). This is a standard retrieval of public data for the purpose of social research and does not involve the execution of remote scripts or binaries.
Audit Metadata