competitor-connection-displacement
Pass
Audited by Gen Agent Trust Hub on Jul 29, 2026
Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
- [PROMPT_INJECTION]: The skill processes untrusted external data via the
new_connectionsinput, creating a surface for indirect prompt injection. Malicious data in the input could attempt to subvert the qualification logic or influence the generated outreach angles. - Ingestion points: External data is ingested via the
new_connectionsinput parameter inSKILL.md. - Boundary markers: No specific delimiters or instructions to ignore embedded commands are present.
- Capability inventory: The skill produces text outputs for prospect qualification and differentiation angles. It does not perform active network or file system operations.
- Sanitization: There are no instructions for sanitizing or validating external input before processing.
- [NO_CODE]: The skill consists entirely of instructional markdown and YAML configuration without accompanying scripts or executable code, which limits the direct execution risk.
Audit Metadata