deep-research
Pass
Audited by Gen Agent Trust Hub on Sep 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill demonstrates secure implementation of command execution. It uses
subprocess.runwith a list of arguments and avoidsshell=True, preventing potential command injection vulnerabilities. - [SAFE]: Secret management follows best practices. The skill retrieves the
OPENAI_API_KEYfrom environment variables or a.envfile rather than using hardcoded credentials. - [INDIRECT_PROMPT_INJECTION]: The skill possesses an inherent attack surface for indirect prompt injection as it ingests and processes untrusted data from the internet via web search.
- Ingestion points: Web search results processed in
assets/deep_research.py. - Boundary markers: The skill does not implement specific boundary markers or 'ignore' instructions for the retrieved web content.
- Capability inventory: Network access for API communication, and local filesystem write access for saving reports (
research_report_*.md) and prompts (research_prompt_*.txt). - Sanitization: No explicit sanitization or filtering is performed on the content retrieved from web searches.
Audit Metadata