← DeepSeek Harness Plugins
DeepSeek Harness PluginManifest Valid

dsh-adaptive-max-tokens

Dynamically adjusts maxTokens output budget as context grows and restores it after compaction.

Developer Tools
0GitHub Stars0ForksUpdated2026-09-27

Install

$ dsh plugin --profile web add github:3289192-bot/dsh-adaptive-max-tokens#v0.1.0

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-adaptive-max-tokens is an adaptive output token budgeting plugin for DeepSeek Harness (DSH). It prevents context window overflow caused by large fixed output reservations by dynamically shrinking maxTokens as input context grows, restoring full output quota after compaction, maintaining calibration across interruptions and resumes, and performing a single budget-only retry for recognized context overflow errors.

Key Capabilities

  • Dynamically adjusts request maxTokens based on estimated remaining context and safety margins
  • Restores max output budget after official compaction triggers
  • Preserves token usage calibration across interrupted replies and conversation resumes
  • Automatically retries once on recognized context overflow errors by reducing output budget
  • Respects adapter and agent routing constraints when switching models

Useful For

  • Preventing LLM provider context rejection during very long conversations
  • Optimizing context space utilization when dealing with massive input prompts and documents
  • Maintaining stable token budgets during interrupted or resumed chat sessions

Who It Fits

  • DeepSeek Harness users working with large context windows and long-form workflows
  • Developers looking to maximize usable input token capacity without hitting premature context limits

Documented Limitations

  • Does not expand the model's actual physical context window or modify compaction algorithms
  • Token usage estimations may introduce variance with large attachments or dynamic tool calls
  • Supports specific DSH versions (@deepseek-ai/dsh 0.1.6-alpha.1 or 0.1.7-rc.2)
  • Only handles recognized context overflow error formats for automatic retries

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
3289192-bot/dsh-adaptive-max-tokens
Registry source
GitHub · dsh-plugin topic
Registry classification
Plugin
Source checked
245c6e9 · 2026-09-27

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.

Repository Activity

GitHub stars
0GitHub stars
Forks
0
Open issues
0
Last commit
2026-09-27
Last release
2026-09-26
For maintainers

Maintaining this plugin?

This listing is generated from public repository data. If you maintain this project, you can review the information and share this listing with your users if you find it useful.

Add to README
Listed on DSHPlugin.app