Skills
skills/louisblythe/sales-skills/win-loss-reason-extraction/Gen Agent Trust Hub

win-loss-reason-extraction

Pass

Audited by Gen Agent Trust Hub on Apr 3, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection through its handling of prospect-provided conversation data.
  • Ingestion points: Untrusted text enters the system via the conversation object processed in SKILL.md within the extract_reasons_with_llm and extract_reasons_from_conversation functions.
  • Boundary markers: The LLM prompt template in SKILL.md (within extract_reasons_with_llm) interpolates raw conversation text directly without the use of delimiters, XML tags, or specific instructions to ignore embedded commands.
  • Capability inventory: The skill is designed to update external CRM records via crm_client.update_opportunity as defined in SKILL.md, potentially allowing injected instructions to corrupt CRM data.
  • Sanitization: There is no evidence of input validation, escaping, or sanitization of message text before it is used in prompt construction.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 3, 2026, 05:46 PM