academic-deep-research

Pass

Audited by Gen Agent Trust Hub on Apr 6, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [SAFE]: Analysis of all skill files, including the extensive SKILL.md methodology, revealed no signs of malicious intent, obfuscation, or unauthorized data access.
  • [NO_CODE]: The skill is composed entirely of markdown instructions and metadata, containing no executable scripts, binaries, or external code dependencies.
  • [EXTERNAL_DOWNLOADS]: The skill facilitates the retrieval of information from the internet using standard web tools to perform literature reviews and data synthesis.
  • [DATA_EXPOSURE]: Includes functions to query the agent's internal memory via memory_search and memory_get, which is used legitimately to maintain continuity with previous research sessions.
  • [COMMAND_EXECUTION]: Orchestrates complex tasks through the sessions_spawn tool to enable parallel research themes by sub-agents.
  • [PROMPT_INJECTION]: The skill processes untrusted data from web fetches; while it lacks explicit software-based boundary markers for this data, its academic protocol—requiring evidence hierarchy, multiple source verification, and identification of contradictions—serves as a robust process-based defense against indirect prompt injection (Ingestion points: web_fetch in SKILL.md; Boundary markers: absent; Capability inventory: web_search, web_fetch, sessions_spawn; Sanitization: absent).
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 6, 2026, 11:30 PM
Security Audit — agent-trust-hub — academic-deep-research