beginner-pm

Pass

Audited by Gen Agent Trust Hub on Apr 1, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes arbitrary natural language from users to drive automated actions. 1. Ingestion points: User project descriptions and clarifying responses. 2. Boundary markers: Absent; no specific delimiters isolate user input from internal instructions. 3. Capability inventory: The skill is authorized to write code, create and configure files, and execute terminal commands. 4. Sanitization: Absent; no explicit validation or sanitization of user-provided content is mentioned.
  • [COMMAND_EXECUTION]: The skill is explicitly instructed to perform autonomous system operations, including running terminal commands and managing file system changes, to fulfill project tasks without manual user intervention.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 1, 2026, 05:35 AM
Security Audit — agent-trust-hub — beginner-pm