wake

Wake captures your terminal sessions including commands, outputs, and git context, storing them in a local SQLite database accessible to Claude Code via MCP. Instead of manually copying terminal output or describing what happened, Claude sees your actual command history and can analyze errors, logs, and build output directly. The tool automatically summarizes command outputs using a local LLM, enabling Claude to quickly understand what occurred without reading full output. Wake integrates seamlessly into your shell workflow—just run `wake shell` to start recording, work normally, then ask Claude to analyze what happened. Features include session search, command history with metadata, tiered output retrieval to optimize context usage, and configurable retention policies.

[{"step":"Install the wake skill in your terminal environment. For most setups, this involves running: `pip install wake-skill` or adding it to your Claude Code extensions directory.","tip":"Ensure your terminal session has access to the project directory and git repository. The skill works best when run from the project root."},{"step":"Execute the wake command with your project path. Example: `wake /path/to/your/project` or simply `wake` if already in the project directory.","tip":"For focused analysis, add flags like `--focus tests/` or `--focus src/auth/`. This narrows the context to specific areas."},{"step":"Copy the generated context summary and paste it into Claude Code or your preferred AI assistant.","tip":"Use the `## Terminal Session Context` header as a delimiter when pasting into long conversations to help the AI understand the boundaries."},{"step":"Ask specific follow-up questions in Claude Code based on the context. Examples: 'Why is the JWT validation failing in the integration tests?' or 'Suggest improvements to the test coverage for the auth module.'","tip":"Reference line numbers or file paths from the context to make your questions more precise. The AI can then analyze the exact code causing issues."},{"step":"Iterate by running additional wake commands after making changes. Example: `wake && ask 'How do these changes affect the test suite?'`","tip":"Combine with terminal tools like `git diff` or `npm run lint` to get even richer context before waking the session."}]