Translation toolkit for DSH agents: sentence segmentation, term and glossary extraction, source-target QA (numbers, units, brackets), tone guides, and a durable workspace translation memory.
Install
# from npm (prebuilt)
dsh plugin --profile web add dsh-plugin-translation
# from GitHub (first run asks for allowBuilds approval — follow the hint, retry)
dsh plugin --profile web add github:863683348/dsh-plugin-translation
Any plugin you install runs third-party code with your own permissions — it can read your files, use your credentials, and reach the network, and tool approvals don’t sandbox it. GitHub-sourced plugins also run build scripts at install time — pnpm blocks those until you allow them, so an install can stop with ERR_PNPM_GIT_DEP_PREPARE_NOT_ALLOWED or ERR_PNPM_IGNORED_BUILDS; dsh prints the exact key to add under allowBuilds in your profile’s pnpm-workspace.yaml, and the install works on the next run. Allowing a build is a trust decision: only install sources you trust, and pin a commit (github:owner/repo#sha).
README
A translation toolkit for DeepSeek Harness agents: chunking, glossary extraction, source–target QA, tone guides, and a durable translation memory. The model translates; the plugin chunks, checks, and remembers.
Compatibility
Tool schemas are validated against the @deepseek-ai/dsh-tools value-schema DSL at plugin load (checked against dsh-tools 0.1.0-rc.6 and 0.1.1-rc.2). Earlier releases used JSON-Schema required at the root of output.schema and closed nested objects without declared properties, which made the host abort the whole profile boot with unsupported JSON schema: schema.required is not supported by the value schema DSL and could reject the tool's own results. Current releases fix both; if an affected version left your DSH unable to start, remove the plugin from the profile (or upgrade) — no data is lost.
Install
dsh plugin --profile <profile> add dsh-plugin-translation
Restart DSH. The translate_kit tool is registered host-wide.
Tool
| action | purpose |
|---|---|
segment |
Split text into numbered, bounded chunks for chunked translation |
glossary |
Extract candidate terms (acronyms, camelCase, domains, numbers+units, emails, URLs) with counts |
check |
Source–target QA: numbers/units, brackets, length ratio, doubled punctuation |
tone |
Register-specific tone guide (formal / colloquial / technical, en / zh) |
memo_get |
Read the translation memory (newest first) |
memo_add |
Save a source→target pair to the memory file in the session workspace |
glossary_get |
Read the durable glossary (newest first) |
glossary_add |
Add or update a source→target glossary term in the workspace glossary file |
glossary_remove |
Remove a glossary term by source |
consistency |
Cross-segment terminology consistency check against the glossary |
tone_get |
Read saved tone guides (filterable by tone and language) |
tone_save |
Save a tone guide for a tone@language pair |
quality |
Heuristic quality score (0-100) with fidelity/fluency breakdown |
Config
All optional, on the composition row's config:
| key | default | meaning |
|---|---|---|
personaSection |
true |
register the translation prompt-guidance section |
sectionOrder |
6 |
prompt section order (persona is 0, ascending) |
memoFile |
.dsh/translation-memo.md |
memory file path (relative to the session workspace; cannot escape it) |
glossaryFile |
.dsh/translation-glossary.md |
glossary file path (relative to the session workspace; cannot escape it) |
maxGlossaryEntries |
200 |
entries kept in the glossary file |
toneFile |
.dsh/translation-tone.md |
tone-memory file path (relative to the session workspace) |
maxToneEntries |
50 |
saved tone guides kept |
maxMemoEntries |
200 |
entries kept in the memory file |
Design
Pure logic (lib/translation.js) has zero DSH/Cordis imports and is unit-tested in isolation; lib/index.js is the thin Cordis plugin. Memory access goes through ctx.fs and every resolved path is containment-checked against the session workspace.
License
MIT
Roadmap
See ROADMAP.md — next five versions (v0.2.0 – v0.6.0): glossary management & consistency, tone memory & quality scoring, side-by-side diff & placeholder protection, conflict detection & effort stats, glossary versioning & batch pipeline.
Links
More in this category
Tencent/WeKnora#dsh-weknora★ 37401
Four read-only tools over a WeKnora knowledge base: list knowledge bases, hybrid passage search, reassemble one document's chunks in order, and WeKnora's own cited RAG or ReAct-agent answer with a resumable session id.
DayuanJiang/next-ai-draw-io#dsh-plugin★ 36932
Draws draw.io diagrams in DeepSeek Harness through the next-ai-drawio MCP server: the link in the reply opens a live preview in the right sidebar, you can edit the diagram by hand, and each AI change is a version you can restore. Node.js 18+; npx downloads the MCP server on first use.
superdesigndev/treg★ 7929
Tool catalog for agents: search ~2,600 external endpoints (SEO and SERP, backlinks, social, people and company enrichment, ad libraries, scraping) by the task you want done, read each one's parameters and per-call price, then call it with the credential injected server-side. Ships the skill plus an MCP row that stays disabled until TREG_TOKEN is set.
T-Auto/dsh-ops★ 5798
Provides bash, PowerShell 7, and high-performance Rust tools for DSH on Windows to reduce token usage.
sixtysevenlf/dsh-blender-plugin★ 2081
Lets a DSH agent drive a local Blender over a direct TCP channel with 15 tools: viewport frames, custom-angle renders, Python execution, headless and job offload, transaction rollback, and a planning layer of 228 ops across 34 families.
yjl9903/AnimeGarden#dsh★ 1400
Anime Garden for DSH: search anime torrent resources from 動漫花園, 蜜柑计划, 萌番组 and ANi inside DSH — via chat, a sidebar panel with fansub and type filters, a seasonal calendar, and an in-app floating player.
Community comments
Comments are public GitHub Discussions. Loading them connects to GitHub and Giscus; a GitHub account is required to post.