Coder/deepseek-harness
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat(dsbench): add SDK evaluation composition

Browse Source
This commit is contained in:
Yichen Jiang
2026-07-17 17:29:38 +08:00
parent a6915745e0
commit a38ff125a7
17 changed files with 365 additions and 29 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
examples/README.md
View File

@@ -21,6 +21,10 @@ Run with: `pnpm run demo:repl` (needs `DEEPSEEK_API_KEY` in the environment or a
Run the Code Mode overlay with `pnpm run demo:code-mode`, or pass `acp` for the ACP example. See the [Code Mode example](coding-agent/README.md#code-mode) for its composition and a sample task.
## dsbench-coding-agent
The unattended SDK composition used by DSBench: JSON-RPC stdio, foreground-only `bash`, `read` / `write` / `edit`, one foreground `subagent`, `todo_write`, JSONL persistence, and compaction. It excludes terminal UI, stdout logging, approvals, skills, and background task controls. See [dsbench-coding-agent/README.md](dsbench-coding-agent/README.md).
## cordis-agent
The **self-referential** demo: the coding spine plus [`@deepseek-ai/dsh-tool-cordis`](../packages/cordis/tool-cordis), whose three tools (`cordis_inspect` / `cordis_mount` / `cordis_unmount`) let the agent inspect the live cordis runtime it runs inside, mount model-written plugins into it (an event listener, a brand-new tool for itself, or a service another mount injects), and dispose them again — all dynamic mounts grouped under one `cordis-dynamic` fiber subtree. The `ctx.fs`/`ctx.web` services ride along provider-only, as the capabilities those plugins build on.