research-knowledge-base

Pass

Audited by Gen Agent Trust Hub on Aug 12, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill retrieves research articles and indexing metadata from https://roy-tong.github.io, which is a domain associated with the skill author. This includes accessing llms.txt, sitemap.xml, and feed.xml.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes and synthesizes data from external web pages. 1. Ingestion points: Content is retrieved from public articles on roy-tong.github.io as described in the Workflow section of SKILL.md. 2. Boundary markers: The instructions do not define explicit delimiters to separate external research content from the agent's instructions, nor do they specify that the agent should ignore commands embedded within the fetched articles. 3. Capability inventory: The skill is limited to information retrieval and synthesis; no capabilities for file system modification, command execution, or network write operations are utilized. 4. Sanitization: The instructions do not describe any mechanisms for validating or sanitizing the content retrieved from the website.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 12, 2026, 07:58 AM
Security Audit — agent-trust-hub — research-knowledge-base