Proactive Agent

# Proactive Agent Makes AI coding agents proactive — anticipate user needs, flag issues, and suggest improvements without waiting to be asked. Anticipate Follow-up Needs — After completing a task, the agent considers what the user might need next and surfaces relevant suggestions automatically. Flag Issues Proactively — Security vulnerabilities, performance bottlenecks, missing error handling, and code smells are surfaced without waiting for explicit prompts. Suggest Improvements — When working on code, the agent notes opportunities for better abstractions, missing tests, documentation gaps, and configuration improvements. Context-Aware Behavior — Uses project context to make relevant suggestions, including CI/CD status, related file reviews, and cross-cutting concerns like logging and i18n. ## How It Works 1 Step 1 Install the skill Add the Proactive Agent skill to your AI coding agent configuration. 2 Step 2 Complete any task The agent finishes the requested work as normal. 3 Step 3 Agent evaluates context After each task, the agent reviews the surrounding code and project context for related issues or opportunities. 4 Step 4 Suggestions surface automatically High-impact, relevant suggestions are presented grouped and prioritized — not as interruptions but as helpful follow-ups. 5 Step 5 You decide Accept, ignore, or defer any suggestion; the agent respects your workflow and learns from your responses. ## Key Features - **Post-task scanning**: Automatically checks for similar bugs, missing tests, or related improvements after completing each task. - **Security and performance alerts**: Proactively flags vulnerabilities, bottlenecks, anti-patterns, and outdated dependencies. - **Grouped suggestions**: Related suggestions are bundled together to avoid noise and reduce interruptions. - **Smart suppression**: Stays silent during time-critical tasks, minimal-output mode, or when suggestions would be disruptive. - **Workflow-aware**: Respects user preferences and adapts based on which suggestions are accepted or rejected. ## Requirements - **No API keys or external tools required**: Works with any AI coding agent out of the box. - **No additional configuration**: Install and the behavior activates immediately. ## Use Cases - **Bug fixes**: After fixing one bug, the agent scans for similar issues elsewhere in the codebase. - **Feature development**: After adding a feature, the agent suggests tests, documentation, and related improvements. - **Refactoring**: After a refactor, the agent surfaces further cleanup and abstraction opportunities. - **Code reviews**: The agent proactively highlights security, performance, and maintainability issues as it works. - **Onboarding**: New contributors benefit from the agent surfacing missing docs, gaps in test coverage, and configuration issues. ## Installation Install via: `npx clawhub@latest install proactive-agent`

インストール

関連スキル