ai-native-dev

Warn

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: MEDIUMPROMPT_INJECTION
Full Analysis
  • Indirect Prompt Injection (MEDIUM): The skill is susceptible to instructions embedded in the project's task board that could manipulate agent behavior.\n
  • Ingestion points: The agent is instructed to read TASKS.md from the project root at the start of every session and periodically during development.\n
  • Boundary markers: None. There are no delimiters or instructions provided to the agent to treat the contents of TASKS.md as untrusted data.\n
  • Capability inventory: The agent has file system write access (specifically to TASKS.md) and is empowered to make autonomous decisions regarding prioritization, effort estimation, and status updates.\n
  • Sanitization: None. The skill explicitly commands the agent to 'not ask permission' and 'just do it' when updating the file, removing the human-in-the-loop safeguard.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 16, 2026, 10:17 AM