pragman-research
Pass
Audited by Gen Agent Trust Hub on Aug 11, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns detected. The skill implements security best practices including egress boundaries and explicit warnings against uploading sensitive data to external providers without approval.\n- [PROMPT_INJECTION]: The skill processes external research data, creating a surface for indirect prompt injection. \n
- Ingestion points: External data retrieval is the primary function, as described in Workflow Steps 4 and 5 in
SKILL.md.\n - Boundary markers: Workflow Step 6 explicitly instructs the agent to keep sourced facts and inferences distinct, and the guardrails specify treating provider output as evidence to verify rather than authority to repeat.\n
- Capability inventory: The skill utilizes browsing and repository access capabilities via the
pragman-routertool.\n - Sanitization: The skill relies on process-based sanitization (egress control) rather than technical filtering of content strings.
Audit Metadata