twitch-research
Pass
Audited by Gen Agent Trust Hub on Oct 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill retrieves user-generated content from Twitch, including stream titles, channel descriptions, and VOD chat logs. This data could contain malicious instructions designed to manipulate the agent's subsequent actions.\n
- Ingestion points: Untrusted data enters the agent context via
scripts/crawlora.shwhen calling endpoints like/twitch/vod-commentsand/twitch/searchdocumented inSKILL.md.\n - Boundary markers: The skill does not provide explicit delimiters or instructions to the agent to treat the fetched Twitch data as untrusted or to ignore embedded instructions.\n
- Capability inventory: The agent can execute shell scripts (
scripts/crawlora.sh) and has general file system and environment access.\n - Sanitization: The skill returns raw JSON data from the API; no sanitization or filtering is performed on the content of Twitch messages or titles before they are presented to the agent.
Audit Metadata