wiki-research-engineering

Pass

Audited by Gen Agent Trust Hub on Aug 17, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection because it ingests and processes untrusted data from external websites and a research queue.
  • Ingestion points: Data enters the context from the web_search tool (using the Brave engine), intelli_extract tool, and by reading the RESEARCH_QUEUE.md file in the vault.
  • Boundary markers: The skill lacks explicit instructions to use delimiters (like XML tags or triple backticks) or specific "ignore embedded instructions" warnings when handling content retrieved from the web.
  • Capability inventory: The skill has permissions to read and write to the local file system within the /app/data/vault/ directory and can perform subsequent web searches based on findings.
  • Sanitization: There is no evidence of content sanitization or filtering to prevent malicious instructions embedded in web pages from influencing the agent's behavior.
  • [COMMAND_EXECUTION]: The skill uses specific search and extraction tools (web_search and intelli_extract) to fulfill its research objectives. While these are part of the intended functionality, they interact with untrusted external environments.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 17, 2026, 11:52 AM
Security Audit — agent-trust-hub — wiki-research-engineering