bearly

When stuck on a specific problem, stop coding, gather context, and ask /deep or /pro for a fresh perspective.

Reframe recurring or wrong-shaped problems into a design where the bug cannot happen.

Run 5 Whys plus /big reframing to trace a symptom to structural root cause before fixing.

Batch operations across files with confidence-based auto-apply. Use for renaming, search-replace, refactoring code, updating text/markdown, migrating terminology, and API migrations. Run `bun tools/refactor.ts --help` for detailed command reference.

Analyze a decision end-to-end: gather context, compare options, score tradeoffs, and present one clear recommendation.

Save active session context, git state, recent commits, and next steps to one tracking issue before compaction or handoff.

Pause implementation for architecture or understanding discussion, checkpoint context, then resume with captured decisions.

Explain current work plainly for a teammate without recent context, defining mechanisms and removing task-specific shorthand.

Multi-LLM research with deep research, second opinions, and multi-model debate. Use for second opinions from other AI models, deep research with web search, building consensus across models, or escaping single-model blindspots.

Batch operations across files with confidence-based auto-apply. Use for renaming, search-replace, refactoring code, updating text/markdown, migrating terminology, and API migrations. Run `bun tools/refactor.ts --help` for detailed command reference.

Multi-LLM research — deep research, second opinions, and multi-model debate. Use when user says /llm.