What problem does it solve?
This Skill provides a comprehensive toolset for MaxTAC research by enabling goal-bounded auditor and debate subagents, facilitating detailed vulnerability analysis and debate processes.
Core Features & Use Cases
- Auditor Subagents: Specialized vulnerability researchers for specific bug classes, mitigations, or security topics, with a 5-step audit flow.
- Debater Subagents: Subagents that review debate topics and vote on ballots, with a 5-step debate flow.
- Parallel or Sequential Execution: The Skill guides the execution of subagents based on system resources and requirements.
- Goal-Bounded Operations: Subagents operate within defined goals, ensuring focused and controlled research processes.
- Use Case: Suppose you have a complex security issue that requires expert analysis and debate. Use this Skill to spawn auditor and debater subagents to analyze the issue from multiple perspectives and determine the best course of action.
Quick Start
Run the audit subagent for the 'Chained Vulnerability Auditor' with a specific context query:
python3 <skill-dir>/scripts/audit-helper.py --root <workspace-root> --prompt-file tmp/<prompt-name>.md --context-query "<hypothesis boundary component>"