yuezengwu/dsh-explain
110Last commit Aug 14, 2026
dsh-explain DSH plugin
dsh-explain maintains one global learning thread across all work sessions, with per-source active explanations. It supports both manual explain requests and autonomous explanations with a configurable budget. Learning state persists across sessions, and users can review explanations even after source sessions are deleted.
How to install the dsh-explain DSH plugin
dsh plugin --profile web add /absolute/path/to/dsh-explainCopying does not run this command. Review the repository and version before installing the dsh-explain DSH plugin.
dsh-explain DSH plugin data source
dsh-explain DSH plugin snapshot date: Aug 16, 2026
discovered
What the dsh-explain DSH plugin can do
- One global learning thread across all work sessions
- Per-source active explanations with feedback loop
- Autonomous explanation requests with configurable budget (default 50 per 24h)
- Manual /explain command and shortcut from selections or assistant answers
- Private ExplainContext summarizing user knowledge and preferences
- Settings UI for auxiliary model, enablement, and budget
- Learning view tab (first-party conversation.view)
- Compaction of auxiliary history when idle or near context limit
- Rephrasing even after source session is deleted
Where the dsh-explain DSH plugin fits
- Learn from completed DSH work sessions without re-reading the entire conversation
- Get explanations of specific concepts or tradeoffs encountered during a session
- Build a persistent knowledge base across multiple work sessions
- Review and rephrase explanations even after the source session is deleted
- Use autonomous requests to automatically generate explanations from new sessions
Who the dsh-explain DSH plugin is for
- DeepSeek Harness users who want to learn from their sessions
- Developers using DSH for complex tasks who need to understand underlying concepts
dsh-explain DSH plugin limitations
- Currently targets DSH 0.1.0-rc.6 only; earlier private-preview packages are not supported
- Autonomous request budget defaults to 50 per 24-hour rolling window; failures and retries count against the budget
dsh-explain DSH plugin: from the repository README
Quoted from the yuezengwu/dsh-explain README, the upstream source of the dsh-explain DSH plugin. Copyright remains with the original authors.
**English** | [简体中文](README.zh-CN.md) # dsh-explain — Learning mode for DSH > ✅ **Status: M6 is implemented entirely in Explain and passes its DSH `0.1.0-rc.6` acceptance gates. No other plugin requires a patch or runtime dependency.** > See the [product requirements](docs/PRD.md) and [technical architecture](docs/ARCHITECTURE.md) for the full design. The detailed design documents are currently written in Chinese. `dsh-explain` turns useful material from multiple [DeepSeek Harness](https://github.com/deepseek-ai/deepseek-harness) work sessions into one private, local-first learning thread. It keeps at most one active explanation per source session and continuously adapts explanations to the user's knowledge, preferences, and learning progress. ## Core behavior - A `$DSH_HOME` contains exactly one local learning thread. Work sessions, resumes, and forks do not copy its state. - Each top-level source session may have one explanation awaiting feedback, or none. While one remains open, that source cannot generate another; other sources may continue independently. - Enter `/explain <learning request>` in the composer of an established or empty session to request an explanation. `/e
Read the full READMERepository license: MIT
dsh-explain DSH plugin questions
How do I install dsh-explain?
Run `npx @deepseek-ai/dsh@0.1.0-rc.6 plugin --profile web add github:yuezengwu/dsh-explain` then `npx @deepseek-ai/dsh@0.1.0-rc.6 web`. Git-hosted plugins build during installation. If pnpm asks for build approval, add the printed 'dsh-explain' entry to the profile's pnpm-workspace.yaml and repeat the install command.
What DSH version does dsh-explain support?
It currently targets DSH 0.1.0-rc.6. Older private-preview package lines are not supported. The source checkout for local development must also match the 0.1.0-rc.6 public API surface.
How does the autonomous request budget work?
The autonomous evaluation defaults to 50 requests per rolling 24-hour window. Usage survives restarts; failures and retries count toward the budget, while manual explanations, rephrases, and compaction do not. You can configure the budget in the settings UI.
Can I use explanations after the source session is deleted?
Yes. The first entry of an explanation stores a restricted source summary of up to 2,000 characters for later rephrasing. Rephrasing still works after the source session is deleted. The summary is not exposed through the learning-view API.
Does the primary model know about Explain?
No. The primary model does not know that Explain exists. Explain does not write to the primary session log, modify the primary model context, or block the primary turn. The ExplainContext is sent only to the auxiliary model and is never injected into the primary agent.