Translation toolkit for DeepSeek Harness agents: sentence segmentation, term/glossary extraction, source-target QA, tone guides, and a durable translation memory in the session workspace
Pending human review — the facts below come from the source repository.
Translation toolkit for DeepSeek Harness agents: sentence segmentation, term/glossary extraction, source-target QA, tone guides, and a durable translation memory in the session workspace
Users who want this capability in DSH; check the source and docs before installing.
dsh plugin --profile web add github:863683348/dsh-plugin-translation No supported platforms were declared.
bash 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 | ## 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) | | 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 MITNot by stars — picked by hand, one at a time.
A visual settings dashboard for DeepSeek Harness: every registered settings namespace — including third-party plugin ones the official UI never covers — rendered as editable, schema-driven forms. Early, but the right direction: stop hand-editing YAML.
A DIY trajectory visualizer for DeepSeek Harness: watch the agent’s plan and execution unfold as a graph instead of scrolling raw logs. Early, but it fills a real gap — understanding at a glance what the agent is doing.
Dispatch DSH work from Claude Code / Codex: spin up in-host DSH agent sessions with per-tier presets, watch native subagent progress, and borrow a multimodal bridge that lends the text-only harness vision and image generation. An orchestration plugin that turns DSH into a backend for other coding agents.