v0.1.1

Full Changelog: v0.1.0...v0.1.1

v0.1.0 — DSH computer-use-linux bridge

First release. dsh-computer-use-linux bridges the
computer-use-linux MCP server
into DeepSeek Harness as native mcp__cul__* tools, with the skill bundled.

Install

# CLI profiles
dsh plugin --profile web add /absolute/path/to/dsh-computer-use-linux

DSH Desktop installs the bundle from its plugin manager. Then restart the
harness: the tool list gains mcp__cul__doctor, mcp__cul__get_app_state,
mcp__cul__list_windows, mcp__cul__click, and the other 14.

Pre-download the desktop binary so the first tool call is fast:

node scripts/install-binary.mjs

What's in it

• 18 desktop tools — diagnostics, window discovery, AT-SPI observation, input
  synthesis, semantic actions, window geometry.
• The computer-use-linux skill, registered additively through
  ctx.skills.register() instead of taking over the shared
  skill-filesystem.bundledSkillDir slot.
• A binary resolver: COMPUTER_USE_LINUX_BIN → plugin cache → npm package →
  PATH → sha256-verified GitHub release download.
• Zero runtime dependencies, no DSH modifications. The bridge is DSH's own
  in-box @deepseek-ai/dsh-mcp-client.

Verified on this commit

• node scripts/check-package.mjs — ok
• node --test — 20 tests, 19 pass, 1 skip (the launcher test needs a cached
  binary, which a fresh runner does not have)
• real MCP handshake through the launcher — computer-use-linux 0.7.13,
  protocol 2024-11-05, 18 tools listed
• CI green on Node 20 and Node 22:
  https://github.com/BruceZhang1993/dsh-computer-use-linux/actions/workflows/ci.yml
• the attached tarball installed into a scratch project: exports resolution,
  package checks, skill loading, and the MCP handshake all pass
  • sha256 275951609c049254b24e5bde573729f2ea619bcf8423540b27e23665bb1aea7e

npm

Not on npm yet. The dsh-computer-use-linux name is free, but the machine that
built this release has no npm credentials. The attached tarball is the
artifact intended for npm; with credentials:

npm publish ./dsh-computer-use-linux-0.1.0.tgz

Requirements

Linux (x86_64 or arm64), Node 20+, a DeepSeek Harness with
@deepseek-ai/dsh-mcp-client, and a desktop session.