Skills
skills/designnotdrum/brain-jar/setup-brainstorm-integration/Gen Agent Trust Hub

setup-brainstorm-integration

Warn

Audited by Gen Agent Trust Hub on Mar 10, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill script uses the cp command to write a configuration file to the user's home directory path ~/.claude/hookify.visual-thinking-brainstorm.local.md. This is a persistence mechanism that modifies the agent's environment and behavioral rules across sessions.
  • [PROMPT_INJECTION]: The installed 'hookify' rule contains persistent instructions that trigger on specific regex patterns within user conversations. When keywords like 'architecture' or 'data model' are detected, the rule injects instructions into the agent's context, overriding default behavior to prompt the user to use the create_diagram tool.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 10, 2026, 02:19 AM