复杂任务范围、步骤与验收计划制定｜简诗 AI：用持久化 Markdown 规划复杂任务
介绍 SkillHub 社区 Skill @user_fccf428b/planning（复杂任务范围、步骤与验收计划制定｜简诗 AI）。它通过创建并管理持久的 Markdown 规划文件，支持结构化任务执行，适用于 create a plan、track progress、start a research project 等场景。文章说明 Skill 定位、核心能力、基于 skillhub CLI 的安装方式（含 Cursor/Claude Code 等 skills 目录），以及典型触发用法与注意事项；无需 API Key，版本 1.0.0。
Read Moredsh-planning-files: Three-File Persistent Planning Plugin
The DeepSeek Harness (DSH) plugin uses a three-file persistence mechanism to prevent task plans, discoveries, and progress from being lost during process restarts or session switches. Maintained by satan9394, the plugin supports managing long-lived planning sessions and must run with the permissions of the current process. It is recommended to review the source code and license before installation. For more details, refer to the GitHub repository or plugin directory.
Read Moretrailmap: A Disk-Persistent Execution Planning Plugin
trailmap is a workflow plugin for the DSH environment that mitigates progress loss in multi-step tasks caused by session interruptions. It persists task states, milestones, and retrospective records to the `.trail/` directory, supporting state-machine transitions, dependency management, cycle detection, and automated retrospectives. The plugin provides a CLI, DSH tools, and a skill interface to ensure tasks are structured and recoverable. Installation requires Node >= 20; pay attention to directory permissions and version control.
Read MoreDeepSeek Harness Long-Horizon Task Suite
DSH's native modules are ill-suited for long-term research or projects. This suite addresses challenges such as task rigidity, heavy reliance on manual intervention, experience loss, and inflated self-assessments through mechanisms including a Mission state machine, Claim Pool task claiming, Blackboard communication, LLM Wiki memory, and forced replanning on failure. Its core features cover long-term autonomous task management, scheduled wake-ups, blind review, and strict validation. The system includes three plugins and two Captain presets, and is compatible with DSH 0.1.0-rc.8 or later. The approach does not hard-code workflows and supports flexible definition of task success criteria and termination policies.
Read MoreDSH-Task-Planner: An Experience-Based Task Planning Plugin for DeepSeek Harness
`dsh-task-planner` is a task planning plugin for DeepSeek Harness, designed to address the issues of lacking historical solution reuse and rule hard-coding mismatches when agents execute complex tasks. This plugin adopts a planning mode with "experiential muscle memory," and its core mechanisms include: 1. **Experience Library Management**: It consolidates historical solutions into editable Markdown files (default path `~/.dsh/planner-lessons`), supporting saving, recalling, and listing experiences, and follows a lifecycle of "draft-validate-discard/skill." 2. **Conditioned-Reflex Planning**: Before a task begins, the plugin first recalls similar historical solutions, and the LLM evaluates their suitability. If suitable, it reuses and improves them; otherwise, it re-plans from scratch, ultimately outputting a step-by-step plan that integrates a capability catalog and risk analysis.
Read More