objection-mining
Pass
Audited by Gen Agent Trust Hub on Jul 29, 2026
Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
- [PROMPT_INJECTION]: The skill contains instructions that explicitly direct the agent to override or ignore user sentiment indicating a lack of interest. Directives such as 'NEVER treat a soft "no" as an exit from follow-up' and the assertion that '"not interested" is not "stop contacting me"' instruct the agent to persist in engagement despite negative user feedback. This behavioral override may conflict with standard AI helpfulness and safety guidelines regarding unwanted persistence or harassment.
- [PROMPT_INJECTION]: The skill exhibits vulnerability to indirect prompt injection due to its processing of external, untrusted data.
- Ingestion points: The skill is designed to mine 'call recordings, reply threads, and lost-deal notes' (SKILL.md).
- Boundary markers: No delimiters or instructions are provided to the agent to distinguish between the content to be analyzed and potential instructions embedded within that content.
- Capability inventory: While the skill itself is instructional, it targets the agent's logic for outreach and sequence management, which could be subverted by malicious input in reply threads to manipulate future agent actions.
- Sanitization: There is no evidence of sanitization, filtering, or validation for the mined transcripts and message threads.
Audit Metadata