Add missing reasoning-effort declarations to custom pi-ai models through DSH settings, enabling the native effort picker and handling models added later.
Install
# from npm (prebuilt)
dsh plugin --profile web add dsh-reasoning-options
# from GitHub (first run asks for allowBuilds approval — follow the hint, retry)
dsh plugin --profile web add github:Scorp1o117/dsh-reasoning-options
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
DSH 0.2.1-rc.2
Supports DSH 0.2.1-rc.2, including explicit prerelease engines.dsh and peer ranges. This host update adds DeepSeek V4.1 Flash 1017 and Search/Fetch APIs; the plugin APIs used here remain unchanged. Existing reasoning settings and the opt-in augmentation policy are preserved.
DSH 0.2.1-rc.1
DSH 0.2.1-rc.1 includes a native editor for custom-model reasoning levels. Automatic augmentation is now opt-in (augmentReasoning: false by default), so turning reasoning off in Models stays off. Existing declarations and provider defaults are retained. OpenCode headers and the revision-checked provider Default helper remain available. On older hosts, opt in to augmentation to restore automatic filling.
Optionally adds seven explicit reasoning levels to custom pi-ai models while retaining DSH's native Default / Provider default choice. Default sends no reasoning effort; it does not turn thinking off at the provider.
Requires DSH 0.2.0-rc.2 or later 0.2.x, on Web and Desktop. Install with the Desktop-provided dsh command (Application → Manage dsh Command):
dsh plugin --profile desktop add dsh-reasoning-options@0.6.0
Use --profile web for Web. Restart DSH after package changes. Open Plugins → Installed → dsh-reasoning-options to configure it.
Default and existing configurations
The plugin no longer inserts reasoning: high on provider routes. New routes without a configured effort show DSH's native Default choice alongside off, minimal, low, medium, high, xhigh and max. The literal string default is never sent to pi-ai or the provider.
For routes configured by an older plugin version, choose Default (omit reasoning effort) in this plugin's provider dropdown. This removes only that route's reasoning setting using a revision-checked unset. Existing configured route defaults are preserved until you change them. If a conversation has an explicit effort selected, select Default again in the conversation's model menu. Default omits effort; the provider may still think according to its own defaults.
Existing custom effort declarations and reasoningEfforts: false are retained. The plugin does not promise that every model supports every explicit level.
Other settings
enabled(true): enable gateway helpers.augmentReasoning(false): opt in to filling missing declarations; this overrides turning reasoning off in the native editor.autoSessionHeader(true): add a missing x-opencode-session header for HTTP(S) opencode.ai and its subdomains.sessionHeaderValue(dsh-session): header value; existing headers are preserved.pollIntervalMs(30000): fallback rescan interval; 0 disables periodic rescans.
Writes use the public settings API, preserve namespace revisions, retry conflicts and run outside HMR transactions. Existing unrelated providers, models and headers are retained. Chinese and English follow the DSH language setting.
Validation
npm test covers declaration reconciliation, preserved route defaults, revision conflicts, HMR scheduling, configuration writes and localization.
Links
More in this category
Ebony-Vinyl/dsh-our-free-model★ 28088
Adds free model lanes to DeepSeek Harness — DeepSeek V4.1 Flash, Kimi K3 and others — with no login, sign-up or API key, plus a local OpenAI-compatible gateway, fallback routing and a usage dashboard.
corrinehu/dsh-workbuddy-connect★ 1450
Brings the models in the WorkBuddy desktop app straight into DeepSeek Harness — zero configuration in the DSH chat.
V1ki/dsh-plugin-subscriptions★ 481
Use ChatGPT (Codex), Claude, and Grok subscriptions as DeepSeek Harness LLM providers, with Settings login, model catalogs, usage, plus image_generate, video_generate, and x_search tools.
Mars-Sea/dsh-commandcode-provider★ 414
Unofficial Command Code LLM provider: registers a `commandcode` route with a live model catalog and reasoning-effort support.
cv-superding/dsh-deepseek-web-login★ 292
Adds a deepseek-web provider that uses chat.deepseek.com web models in DSH, with browser login capture, PoW request signing, SSE streaming, and prompting-based tool calls.
FishBottle7/opencode2dsh★ 284
Exposes OpenCode Zen free models to DeepSeek Harness, with no API key.
Community comments
Comments are public GitHub Discussions. Loading them connects to GitHub and Giscus; a GitHub account is required to post.