AgentSkillsCN

red-team-adversary

通过在沙箱环境中主动尝试利用已知漏洞，开展积极的安全“攻防演练”，并在静态扫描之外进一步验证威胁的真实存在性。

SKILL.md
--- frontmatter
name: red-team-adversary
description: Performs active security "war gaming" by attempting to exploit identified vulnerabilities in a sandbox. Validates threat reality beyond static scans.
status: implemented

Red-Team Adversary

This skill takes a proactive, offensive stance on security to ensure defenses are truly effective.

Capabilities

1. Controlled Exploitation

  • •Attempts to exploit vulnerabilities found by security-scanner within a local sandbox or staging environment.
  • •Provides "Proof of Concept" (PoC) for critical bugs to demonstrate real impact.

2. Resilience Testing

  • •Simulates common attack vectors (DDoS, SQLi, Credential Stuffing) to test the robustness of the crisis-manager and disaster-recovery-planner.

Usage

  • •"Perform a red-team audit on the authentication module and try to bypass it."
  • •"Verify if the SQLi vulnerability found yesterday is actually exploitable in our current setup."

Knowledge Protocol

  • •This skill adheres to the knowledge/orchestration/knowledge-protocol.md. It automatically integrates Public, Confidential (Company/Client), and Personal knowledge tiers, prioritizing the most specific secrets while ensuring no leaks to public outputs.