argo-deep-research
Pass
Audited by Gen Agent Trust Hub on Nov 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and synthesize multi-source evidence, academic literature, and search results. This external data constitutes a surface for indirect prompt injection, where malicious instructions embedded in web pages or papers could influence the agent.
- Ingestion points: External data enters the context via
scripts/research.py, theargo searchcommand, and JSON data piped intoscripts/evidence.py. - Boundary markers: No explicit delimiters or instructions to ignore embedded content are present in the SKILL.md documentation.
- Capability inventory: The skill executes local Python scripts (
scripts/research.py,scripts/evidence.py) and theargocommand-line tool. - Sanitization: There is no evidence of input validation or content filtering for the processed research data within the provided instructions.
- [COMMAND_EXECUTION]: The skill uses local command-line execution for its primary functions.
- Evidence: Usage of
python3 scripts/research.py,argo search, andpython3 scripts/evidence.pyfor performing its core research tasks.
Audit Metadata