The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses uv run to execute several internal Python scripts for session management and configuration resolution.
Evidence: SKILL.md invokes resolve_customization.py, resolve_config.py, memlog.py, and brain.py to handle the brainstorming workflow and library management.
[COMMAND_EXECUTION]: The agent is instructed to use system commands to open generated HTML files in the user's default browser.
Evidence: SKILL.md specifies using open, xdg-open, or start to display the brain-selector.html and final session artifacts.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it processes untrusted user input and session logs to generate ideas and artifacts.
Ingestion points: User-provided brainstorming topics, goals, and existing session history stored in .memlog.md files.
Boundary markers: The skill uses the memlog.py utility to maintain a structured record and performs a deliberate 'synthesis' step to process collected ideas.
Capability inventory: The skill can execute local scripts, write files to the workspace, and open URLs/files in the browser.
Sanitization: The brain.py script uses html.escape when generating the selector page to prevent script injection in the rendered HTML output.

bmad-brainstorming