Scan Report
5 /100
lobster-coordinator
三层多Agent协调器（参考Claude Code架构），支持AgentTool单任务委托、Swarm团队协作、Coordinator模式全局调度。当任务复杂需要多Agent协作时自动激活。
Pure documentation skill describing a multi-agent coordination architecture with no executable code or suspicious behavior.
Safe to install
This is a low-risk documentation-only skill. No action required.
| Resource | Declared | Inferred | Status | Evidence |
|---|---|---|---|---|
| Filesystem | NONE | NONE | — | N/A - documentation only |
| Network | NONE | NONE | — | N/A - documentation only |
| Shell | NONE | NONE | — | N/A - documentation only |
| Environment | NONE | NONE | — | N/A - documentation only |
| Skill Invoke | READ | READ | ✓ Aligned | SKILL.md describes platform-level orchestration |
| Clipboard | NONE | NONE | — | N/A - documentation only |
| Browser | NONE | NONE | — | N/A - documentation only |
| Database | NONE | NONE | — | N/A - documentation only |
File Tree
1 files · 6.2 KB · 212 lines Markdown 1f · 212L
└─
SKILL.md
Markdown
Security Positives
✓ Pure documentation skill with no executable code
✓ No suspicious patterns like base64 encoding, eval(), or obfuscation
✓ No credential harvesting or environment variable iteration
✓ No remote script execution patterns (curl|bash, wget|sh)
✓ Describes legitimate multi-agent coordination architecture
✓ No dependencies or third-party packages to audit
✓ No sensitive file paths accessed (~/.ssh, ~/.aws, .env)
✓ References Claude Code architecture - standard reference implementation