| 123456789101112131415161718192021222324252627282930313233343536 |
- # Code Mode over the shipped TUI: the model stops calling one tool per turn and
- # instead writes TypeScript programs for a single `run_code` tool, executed in a
- # worker thread.
- #
- # This file is a `--config` OVERLAY, not a tree: `dsh --config
- # examples/code-mode/cordis.yml` includes the shared base, applies the TUI
- # overlay, then applies these patches, all at one include level. Patches never
- # cross an include boundary, so an overlay must stay a patch list rather than a
- # file that includes another config.
- #
- # A patch replaces the targeted row's whole `config`, so each row restates every
- # key it owns.
- - id: tools
- config:
- mode: code
- - id: system-prompt
- config:
- persona: |
- You are a coding agent powered by the {{model}} model.
- You work by writing TypeScript programs for run_code: batch related
- tool work into one program, loop and branch where it helps, and print
- or return ONLY the findings that matter.
- - id: tui
- config:
- sessionId: !!js configuredAgentIdentities?.main?.id ?? 'main'
- welcome: 'TUI Code Mode ready. Give it a multi-tool task.'
- showReasoning: true
- maxToolOutputLines: 6
- # The worker thread `run_code` executes in; the base mounts no code runtime.
- - insert:
- - id: code-runtime
- name: '@deepseek-ai/dsh-code-runtime-worker'
|
