Skills
Navigate
skills/giggle-official/skills/giggle-minimax-h3-gen/Gen Agent Trust Hub

giggle-minimax-h3-gen

Fail

Audited by Gen Agent Trust Hub on Sep 13, 2026

Risk Level: HIGHPERSISTENCEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PERSISTENCE]: The setup_api_key function in scripts/generation_api.py modifies user shell configuration files, including ~/.bashrc, ~/.zshrc, ~/.bash_profile, and ~/.profile. It appends an export GIGGLE_API_KEY command to these files to ensure the API key persists across sessions.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and optimize untrusted user input to create video generation prompts.
  • Ingestion points: User descriptions provided during the prompt optimization phase in SKILL.md.
  • Boundary markers: None explicitly enforced to separate user-provided content from the agent's instructions.
  • Capability inventory: The skill uses requests in scripts/generation_api.py to send data to the giggle.pro API and requires python3 for execution.
  • Sanitization: The skill does not implement validation or sanitization of the user input before incorporating it into the final optimized prompt.
  • [EXTERNAL_DOWNLOADS]: The skill utilizes the requests library to communicate with https://giggle.pro, which is the official domain of the skill's author, giggle-official.
  • [COMMAND_EXECUTION]: The skill requires the execution of python3 to run the scripts/generation_api.py script for API configuration and video task submission.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Sep 13, 2026, 03:56 AM
Security Audit — agent-trust-hub — giggle-minimax-h3-gen