菜单
Check installed plugins for conflicts with super-agent-skills and suggest complementary MCP servers. Use to optimize your plugin setup.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
| name | plugin-audit |
| description | Check installed plugins for conflicts with super-agent-skills and suggest complementary MCP servers. Use to optimize your plugin setup. |
| phase | meta |
| produces | ["audit-report"] |
Check your Claude Code environment for plugin conflicts, identify complements, and suggest MCP servers that would enhance your workflow.
This is advisory only. The audit reports findings and recommendations. It does NOT make any changes — you decide what to act on.
Announce at start: "I'm running a plugin audit to check for conflicts and suggest improvements."
Run /plugin or check settings to identify installed plugins. Flag any that overlap with super-agent-skills:
Known conflicts:
| Plugin | Conflict | Recommendation |
|---|---|---|
superpowers | Overlaps entirely — orchestration chain, TDD, debugging, code review | Disable. super-agent-skills replaces it completely. |
agent-skills | Overlaps entirely — engineering standards, domain skills | Disable. super-agent-skills replaces it completely. |
feature-dev | Duplicates code-explorer, code-architect, code-reviewer agents. Its /feature-dev command bypasses /superthink. | Disable OR pick one workflow. super-agent-skills integrates equivalent agents natively into brainstorming and writing-plans. |
How to check:
# List installed plugins (via Claude Code)
# Check /plugin → Installed tab
# Or check ~/.claude/settings.json for enabledPlugins
Report format:
⚠️ CONFLICTS (recommend disabling):
- [plugin]@[marketplace] — [reason]
→ Disable: /plugin disable [plugin]@[marketplace]
If no conflicts found:
✅ No conflicting plugins found.
Check for plugins and MCP servers that enhance super-agent-skills:
Known complements:
| Plugin/MCP | Enhances | Benefit |
|---|---|---|
context7 (plugin or MCP) | source-driven-development | Real-time documentation |
github (plugin) | All skills | Issue/PR integration |
sentry (MCP) | systematic-debugging | Production error data |
| Browser MCP | browser-testing-with-devtools | Automated browser testing |
| PostgreSQL MCP | api-and-interface-design | Schema exploration |
frontend-design (plugin) | frontend-ui-engineering | Design system guidance |
code-review (plugin) | requesting-code-review | Additional review tooling |
| TypeScript LSP (plugin) | All TypeScript work | Code intelligence |
caveman (plugin) | All skills | Token compression — drops filler/articles, preserves technical substance. Reduces output tokens ~75%. |
How to check MCP:
# Check configured MCP servers
# /mcp command in Claude Code
# Or check .mcp.json files
Report format:
✅ COMPLEMENTS (already installed):
- [name] — enhances [skill]
Based on the project type, suggest MCP servers that would help:
Detection logic:
package.json → suggest Context7, TypeScript LSP.py files → suggest Python LSP.env with SENTRY_DSN → suggest Sentry MCPReport format:
💡 SUGGESTED (not installed):
- [name] — [what it does for you]
→ Install: [exact command]
If nothing to suggest:
✅ Your setup looks complete for this project.
Combine all findings into a single report:
# Plugin Audit Report
## ⚠️ Conflicts
[list or "None found"]
## ✅ Complements Installed
[list or "None detected"]
## 💡 Suggested Additions
[list or "Setup looks complete"]
## Actions
[Numbered list of recommended actions, most important first]
| Thought | Reality |
|---|---|
| "I'll check my plugins later" | Conflicting plugins cause subtle issues (duplicate skills, conflicting advice). Check now. |
| "I don't need MCP servers" | You don't need them, but they make skills significantly more effective. A 5-minute setup pays off across every session. |
| "More plugins is better" | Conflicting plugins are worse than fewer plugins. Quality over quantity. |
After running the audit:
Initialize or audit a project's CLAUDE.md against best-practice criteria. Use on first-time setup ('init', 'set up project', 'first time in this repo'), when scanning a project to create CLAUDE.md, or when refreshing an existing CLAUDE.md against the current state of the codebase. Replaces Claude Code's default /init.
Use when starting any conversation - establishes how to find and use skills, requiring Skill tool invocation before ANY response including clarifying questions
Optimizes agent context setup. Use when starting a new session, when agent output quality degrades, when switching between tasks, or when you need to configure rules files and context for a project.
Use when completing tasks, implementing major features, or before merging to verify work meets requirements
Use when receiving code review feedback, before implementing suggestions, especially if feedback seems unclear or technically questionable - requires technical rigor and verification, not performative agreement or blind implementation
Hardens code against vulnerabilities. Use when handling user input, authentication, data storage, or external integrations. Use when building any feature that accepts untrusted data, manages user sessions, or interacts with third-party services.