DeepSeek Harness Plugins
DeepSeek Harness PluginManifest Valid

dsh-prime-agent

Persistent TypeScript REPL and RLM control plane for DeepSeek Harness Code Mode.

Terminal & TUIDeveloper ToolsSecurity & PolicySkills & WorkflowsRemote Execution
5GitHub Stars0ForksUpdated2026-09-09

Install

$ dsh plugin --profile web add ./dsh-prime-agent https://github.com/yoke233/dsh-tool-monitor/archive/1e0f2cc14b4ddbc49c2e3cb2c2a7913c80b3083f.tar.gz

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-prime-agent is a Prime Agent-inspired persistent RLM control plane for DeepSeek Harness. In sessions using the Prime preset, it exposes a single model-visible tool (repl) executing inside a persistent TypeScript REPL worker. Declared top-level variables, functions, and objects survive across cells in the same session, while standard DSH capabilities (tools, agents, jobs) are injected as preloaded bindings inside the execution realm.

Key Capabilities

  • Persistent TypeScript REPL where top-level variables and functions persist across execution cells
  • Preloaded bindings for standard DSH tools, child subagents, and background jobs within the cell environment
  • Integrated apply_patch helper matching OpenAI Codex patch semantics across multiple hunks and files
  • Model prompt minimization exposing only the repl tool to prevent routing confusion
  • Single-slot retention mechanism capturing the latest non-undefined completion accessible via $_
  • Refine skill and slash command (/refine) for continual learning and session adaptation
  • Parent-process lifecycle monitoring and graceful realm/worker cleanup

Useful For

  • Iterative TypeScript and JavaScript development where state and intermediate data must persist across multiple turns
  • Complex code transformations, multi-file patching, and inspection using a programmatic REPL interface
  • Long-running autonomous agent workflows orchestrating subagents and background terminal jobs inside code

Who It Fits

  • DeepSeek Harness power users wanting an RLM (Recursive Language Modeling) notebook-style execution environment
  • Developers seeking persistent TypeScript execution state instead of stateless one-shot tool calls
  • Autonomous agent builders leveraging multi-turn programmatic control loops

Documented Limitations

  • Realm namespaces are live-only in memory; timeouts, cancellations, or hard worker kills reset transient variables
  • Does not support running detached as a background daemon; requires an active foreground dsh lifecycle
  • Requires DeepSeek Harness runtime with Cordis patch compatibility

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.

No automated security signals have been published yet.

Source & Registry Notes

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

Source repository
yoke233/dsh-prime-agent
Registry source
GitHub · dsh-plugin topic
Registry classification
Plugin Bundle
Source checked
cf971c0 · 2026-09-16

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
5GitHub stars
Forks
0
Open issues
0
Last commit
2026-09-09
Last release
2026-09-04
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