79 results
LLM auto-review approval answerer for DeepSeek Harness — decides sandbox escalations without a human prompt via a deterministic filter and a clean-context LLM safety review. REQUIRES a patched harness core (see core-patches/)
DSH plugin: URL parse/clean/decode/encode/expand tool (url_tool), zero-dependency pure ESM.
Context compression for small-context models on the DeepSeek Harness Web GUI: one click condenses the whole conversation into a few sentences, opens a fresh session in the same workspace, and re-injects the summary as the first message so a small model keeps working with a clean, compact context.
家电清洁
年前大扫除
Disk-usage audit for DeepSeek Harness (dsh) data directories: total size, per-directory breakdown, largest files, and oversized-file warnings (session logs can hit hundreds of MB). Zero runtime dependencies, read-only. CLI + agent-callable disk_audit tool.
架构模式：Clean Architecture、Hexagonal、DDD 战术模式、依赖规则。受 wshobson/agents（38k★ MIT）启发。
Session management plugin for DeepSeek Harness (dsh): list / delete / archive & restore sessions, with a built-in "Archived" group and a delete confirm dialog (real delete, not trash). 为 DeepSeek Harness 提供会话管理：列表 / 真删除（二次确认）/ 归档分组与恢复。
DSH Web 三栏工作台：对话流只保留 AI 输出（工具调用/思考移入右侧面板），左侧工作区文件树带 git 状态徽标。Three-column workbench for the DSH web GUI: clean conversation flow, workspace file tree with git badges, and a live tool-call/think panel.
历史消息记录：识别每条用户消息（可展开回复/工具子记录），支持定位、复制、↑↓回溯、按条删除与原位恢复（内置节点导航条，删除后工具调用卡片无遗留）
Export a DSH session log as a clean chat log: stream fragments folded into normal messages, every other event kept verbatim. /chat command + browser-download button.
Clean Architecture 参考库(agent 审查基线): 8 仓钉版 FTS5 搜索 + 断言清单 + 自动审查台账 + 只读观察窗
Safe start/stop companion for dsh web: a bottom-right power dock in the Web UI that shuts the host down cleanly (response flush, port release, exit code 0) and reports the running dsh version.
Static and runtime security guard for the DeepSeek Harness: scans plugins and workspaces for malicious code, context injection and token waste (block/warn/clean), intercepts dangerous runtime tool calls and prompt steps, and exposes /scan, plugin_scan, the scan web panel and a user-managed allowlist. Fully static analysis — never executes scanned code.
Keeps long coding-agent sessions lean: wrap work in named tasks and, when one is done, fold its whole span into a short titled summary. The conversation stays readable, context costs stay low, and every fold's original content can be read back on demand. For DeepSeek Harness.
dsh desktop shell: wrap the running dsh web (Harness) into a real desktop client — auto-launch, protocol file, shortcut registration & cleanup, autostart setting
Archived-session tools for DeepSeek Harness: restore, safely delete, clean leftovers, or continue with a new preset.
A /rewind picker for DeepSeek Harness: scroll to an earlier message and continue from a clean forked session, trimming all later context.
DSH profile bundle: Settings pages to list current/residual sessions and archived documents (ID, date, summary) and delete them with double confirmation.
Local clean-Markdown full or selective-turn export for the current DeepSeek Harness Web conversation.
DeepSeek bills peak hours at 2x, and peak is 09-12 and 14-18 Beijing time. npx dsh-lean audit shows what your session paid and what it costs off-peak. DeepSeek 峰时按 2 倍计费，峰时正好是上班时间。
DeepSeek Harness tools for external opportunity discovery, unmet-need research, pain-point signals, opportunity mapping and validation experiments.
简洁直观的轨迹视图 — 按用户消息分组，显示 Token 消耗，差异化展示用户/模型/工具，支持加载更早记录和内容搜索，适配所有主题
清洁频率