ai-research-refine

Pass

Audited by Gen Agent Trust Hub on Jun 13, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill facilitates research and data analysis by communicating with the vendor's official API at ai-skills.ai. This is standard functionality for this type of service and does not involve any unauthorized data exfiltration or suspicious network activity.
  • [PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data through parameters like materialUrl, materialFile, and materialText. This inherent functionality creates a surface for indirect prompt injection, where adversarial content within user-supplied research materials could attempt to manipulate the agent's analysis or output.
  • Ingestion points: materialUrl, materialFile, and materialText parameters defined in references/form-schema.json.
  • Boundary markers: No explicit delimiters or instruction-ignore markers are specified for the interpolation of these data sources.
  • Capability inventory: The agent's capabilities are focused on text generation and analysis; the skill itself does not contain local scripts for arbitrary command execution or system modification.
  • Sanitization: Input for materialFile is restricted to spreadsheet and data formats (.xlsx, .xls, .csv, .tsv), which mitigates the risk of ingesting executable scripts or malicious binary formats.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 13, 2026, 03:31 AM
Security Audit — agent-trust-hub — ai-research-refine