researchers-primary-source
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted data from a wide variety of external sources (social media, blogs, forums, IRC logs, etc.) using
WebFetchandWebSearch. This creates an attack surface where malicious instructions embedded in these primary sources could potentially be interpreted and followed by the agent. - Ingestion points: Data enters the agent context from arbitrary URLs found during research across platforms like Twitter, Reddit, Hacker News, and personal blogs (detailed in both
SKILL.mdandDIRECT_SOURCES.md). - Boundary markers: While the skill defines a structured output format for research findings, it does not provide the agent with explicit instructions or delimiters to isolate and ignore potentially malicious commands contained within the fetched content.
- Capability inventory: The skill is granted extensive capabilities, including web access (
WebFetch,WebSearch) and file system access (Read,Edit,Write,Grep,Glob), which increases the potential impact of an injection. - Sanitization: No sanitization, filtering, or validation mechanisms are specified for the content retrieved from the web before it is presented to the agent for analysis.
Audit Metadata