skills/tanweai/pua/mama/Gen Agent Trust Hub

mama

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEPROMPT_INJECTIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill instructs the agent to adopt a specific high-pressure persona ('Chinese Mom Nagging Mode') to influence output style and behavior. It utilizes psychological manipulation techniques, such as emotional blackmail and social comparison, to pressure the agent into thorough task execution and debugging.
  • [INDIRECT_PROMPT_INJECTION]: The skill defines a potential attack surface by directing the agent to dynamically ingest behavioral constraints from a external file found via searching.
  • Ingestion points: The agent is explicitly instructed to use the Glob tool to search for the path **/pua/skills/pua/SKILL.md and read its content upon loading the skill.
  • Boundary markers: The instructions lack delimiters or warnings to treat the ingested file as untrusted data or to prevent it from overriding existing safety boundaries.
  • Capability inventory: The skill references the agent's access to searching, file reading, and command execution tools as part of its 'house rules.'
  • Sanitization: No validation, sanitization, or integrity checks are performed on the content of the target file before it is integrated into the agent's core behavior.
  • [COMMAND_EXECUTION]: The skill encourages the agent to make extensive use of its command execution and troubleshooting tools to ensure tasks are completed without human intervention.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 06:39 AM
Security Audit — agent-trust-hub — mama