ln-201-opportunity-discoverer

Pass

Audited by Gen Agent Trust Hub on Apr 25, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill includes a fallback mechanism to fetch missing shared reference documentation from the author's official repository at https://raw.githubusercontent.com/levnikolaevich/claude-code-skills/master/.
  • [COMMAND_EXECUTION]: The skill utilizes shell commands for environment setup and reporting, specifically mkdir -p to create research directories and standard file write operations for documentation.
  • [PROMPT_INJECTION]: The skill processes external data retrieved via web searches to evaluate product ideas, which represents a potential surface for indirect prompt injection. The skill implements a structured 'KILL funnel' logic that helps focus the analysis on specific metrics, though it lacks explicit delimiters or sanitization for search content. Ingestion points: Web search results (Filter 1-4). Boundary markers: None. Capability inventory: WebSearch, Write, Bash. Sanitization: None mentioned.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 25, 2026, 07:28 PM