CommandCodeAI / CommandCodeAI/command-code
Feature Request: `/btw` — Interrupt & Ask Mid-Operation
Nobody has claimed this yet.
- Dominant language
- No language data
- Stars
- 4k
- Forks
- 350
- PR merge metrics
- No merged PRs in 30d
Description
What is /btw?
Both Claude CLI and QWEN CLI have a /btw feature. While the model is mid-operation (generating code, reviewing files, exploring a codebase), the user can type /btw to interrupt the model and ask a question about what it's cur$
The model then:
- Pauses its current output
- Answers the user's question
- Resumes (or adjusts) what it was doing
Use Cases
- Model is writing a long refactor, user realizes a clarification is needed →
/btw"wait, does this handle the edge case where X is null?" - Model is exploring a codebase →
/btw"also check the auth middleware" - Model is reviewing code →
/btw"what does that function actually do?"
Why Extensions Can't Add This
None of Command Code's user-extension mechanisms can implement /btw:
| Mechanism | Why It Can't |
|---|---|
| Hooks | Only fire before/after tool calls, not during model output |
| Skills | Instruction sets, can't modify the stdin/stdout event loop |
| Agents | Behavior configs, can't add new interaction modes |
| MCP | External tool connections, unrelated to chat loop |
| Slash commands | No user-defined slash commands supported |
/btw requires changes to the core conversation event loop: listening for a special input during model output streaming, interrupting the stream, injecting context, and resuming. This is CLI-internal plumbing.
Request
Please add a /btw (or similar interrupt-and-ask) slash command that lets users pause the model mid-operation to ask clarifying questions.
Contributor guide
No contributing guide indexed for this repository
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start at the core conversation event loop and model output streaming path described in the issue, and trace how user input is handled during generation. Determine the interruption, question injection, and resume behavior needed; done means a user can invoke /btw mid-operation, receive an answer, and resume or adjust the original task.
Written by the indexing model from the issue text.
Assessment
- Domain
- ai, cli
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Quiet
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100