DeepSeek Harness PluginManifest Valid

dsh-thinking-effort

Configurable reasoning effort levels and subagent defaults for DSH llm-pi-ai models.

Developer Tools

15GitHub Stars1ForksUpdated2026-08-25

Install

$ dsh plugin --profile <profile> add @hytime/dsh-thinking-effort

Plugin Overview & Capabilities

AI-assisted organization based on the public repository snapshot. The content must be grounded in source evidence and does not replace compatibility or security verification.

source-grounded

dsh-thinking-effort is a DeepSeek Harness plugin that provides configurable reasoning effort levels for hand-declared llm-pi-ai models. It supplies a dedicated settings UI to configure reasoning levels per model, maps DSH reasoning levels to custom gateway parameters, and defines default reasoning effort levels for subagents without overriding explicit values.

Key Capabilities

Adds default reasoning levels (off, high, max) to undeclared llm-pi-ai models
Provides a per-model settings UI to customize reasoning levels and gateway values
Maps standard DSH effort levels to custom backend or gateway parameters (e.g. mapping high to ultra)
Configures global default reasoning effort levels for subagent requests
Integrates multilingual settings UI with Chinese, English, Japanese, and Korean dictionary support

Useful For

Enabling reasoning effort selectors in Composer for custom third-party models declared via llm-pi-ai
Mapping DSH reasoning parameters to non-standard upstream API thinking parameters
Standardizing thinking effort defaults across subagent invocations in multi-agent workflows

Who It Fits

DSH users deploying custom or third-party reasoning models through llm-pi-ai
Developers needing granular control over gateway-specific thinking effort parameters
Users running subagent workflows that require predictable default reasoning behaviors

Documented Limitations

Japanese and Korean UI languages require core DSH runtime locale support or forked locale definitions
Host configuration changes require restarting DSH
Configured subagent levels must be supported by the target model to avoid upstream UNSUPPORTED_REASONING_EFFORT errors
Omitting the reasoning parameter may rely on gateway-default behavior depending on protocol implementation

DSH Compatibility

Version-specific runtime evidence collected by DSH Plugin. A missing result means we have not tested that combination yet.

Not tested yetNo runtime compatibility tests have been published yet.

Security Signals

Objective signals discovered from package metadata and source inspection. These are not a guarantee that a plugin is safe.

Dependency Counts

package.json declares 0 runtime, 0 development, 1 peer, and 0 optional dependencies.

info

Dsh Bundle Declared

package.json declares DSH bundle metadata.

info

License Declared

GitHub reports the repository license as MIT.

info

Package Manifest Available

A root package.json was captured and can be inspected by the registry.

info

Source Available

Public GitHub source metadata is available for this registry snapshot.

info

Source & Registry Notes

Public provenance, Registry classification, and the latest source check for this entry, kept separate from runtime verification.

Source repository: hytime/dsh-thinking-effort
Registry source: GitHub · dsh-plugin topic
Registry classification: Plugin
Source checked: 38f5410 · 2026-08-28

This project is independently indexed from public source information. DSH Plugin is not affiliated with DeepSeek or the plugin author. Always check the author repository before installation.