crw-research
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from an external source (the fastCRW Research API) and incorporates it into the agent's conversational context, creating an attack surface for indirect prompt injection.
- Ingestion points: Paper metadata, including primary IDs, titles, and abstracts, are fetched from
api.fastcrw.comand processed by the agent in SKILL.md. - Boundary markers: No specific delimiters or instructions are provided to help the agent distinguish between legitimate research data and potentially malicious instructions embedded within the fetched results.
- Capability inventory: The skill uses
Bash(curl),Bash(jq), andReadtools, which allow it to fetch data and read local files. - Sanitization: There is no evidence of filtering or sanitization of the content retrieved from the API before it is presented to the agent.
- [COMMAND_EXECUTION]: The skill uses
curlto perform HTTP GET requests to the research API andjqto parse and format the JSON response. - [EXTERNAL_DOWNLOADS]: The skill connects to the external domain
api.fastcrw.comto retrieve paper data. As this is a vendor-owned domain associated with the skill's author, the connection itself is documented neutrally.
Audit Metadata