plur-ai/dsh-plugin0

@plur-ai/dsh

PLUR memory for DeepSeek Harness — engrams injected straight into the prompt, no tool call required

AI 分析

核心用途是为 DSH 智能体提供免工具调用的记忆注入功能。适合希望智能体能自动检测并记住用户纠错、跨会话保持上下文记忆的 DSH 用户。

パッケージ
@plur-ai/dsh
バージョン
0.1.0
ライセンス
Apache-2.0
最終更新
2026/08/15

インストール

$npx -p @deepseek-ai/dsh dsh plugin --profile web add github:plur-ai/dsh-plugin

ドキュメント

README 全文を読む ↗

Configuration

All settings live under the plur namespace in $DSH_HOME/settings.yaml (usually ~/.dsh/settings.yaml).

SettingDefaultMeaning
path~/.plurStore location
scopederivedWhich memory scope this harness may read and write. Omitted, it derives per workspace
injectionModecontentcontent injects memories; off disables injection
injectionBudget2000Approximate token ceiling for the injected block
refreshIntervalMs0Floor between recalls; 0 means once per turn
autoLearntrueDetect corrections in your messages and store them
autoCapturetrueRecord an episode summary at turn end
timeoutMs5000Hard bound on any single memory call
viewerEnabledtrueRegister the /plur-memory command
includeGlobaltrueWhether global engrams accompany the workspace scope

Reranking is configured through PLUR itself, not here: set the PLUR_RERANKER environment variable. It runs in the harness's own process and bge-reranker-v2-m3 peaks around 2GB RSS, where a native OOM cannot be caught by a JavaScript try/catch and would take your agent down with it. Leave it unset for interactive use; enable it for local batch work where a crash costs you nothing.