Reviews Deep Agents code for common configuration mistakes and runtime gotchas. It checks for the critical stuff like using interrupt_on without a checkpointer, forgetting the store parameter with StoreBackend, or expecting files to persist across threads with the default StateBackend. Also catches backend misconfiguration, subagent setup errors, and filesystem tool path issues. The anti-confabulation gate is interesting: it forces the reviewer to quote the actual code before flagging anything, which should reduce false positives from LLMs hallucinating problems. Use this when you're working with create_deep_agent, custom backends, or the human-in-the-loop patterns and want a second pair of eyes on the wiring.
npx skills add https://github.com/existential-birds/beagle --skill deepagents-code-review