237 results
Evidence-driven Feishu/Lark document writing skill for DeepSeek Harness (dsh) — turn Feishu wiki/docs/minutes/Base/chats & local files into structured docs with conclusions, evidence & actions. 飞书文档证据化编写技能插件：从业务资料到可验证、结构化、可执行的飞书成稿（战略方案/经营分析/GTM/复盘/PRD/管理汇报等）。
DeepSeek Harness 的 Skills 管理器：在 设置 → 插件 → Skills 标签页中发现 GitHub 上的 Agent Skills 仓库（SKILL.md 技能包，topic:agent-skills / dsh-skill / claude-skills），查看详情后一键安装 / 移除。技能装入 $DSH_HOME/skills，skill-filesystem 热加载，无需重启。支持 GitHub 代理。
Evidence packaging, Skill evaluation, and controlled trajectory comparison for DeepSeek Harness sessions
发布上线：发布检查清单、灰度、回滚预案、上线后监控。受 addyosmani/agent-skills（88k★ MIT）启发。
DSH web plugin: Skill Cockpit panel in Settings. See all skills (system/project/pool) with Chinese notes, search, enable/disable/delete/import, local test-fire scoring, and per-skill usage stats to spot zombie skills.
源码驱动开发：每个框架决策都要官方文档背书，不凭记忆实现，验证并引用来源。受 addyosmani/agent-skills（88k★ MIT）启发。
AutoMemory for DSH：会话级记忆（收尾提醒 + 软上限检查 + 记忆索引维护），配置面板在 设置→插件。
OpenGame browser-game planning and building Skill for DeepSeek Harness
电商素材工坊（中英双语）。用户需要生成电商主图、详情图、场景图、产品图合成、品牌叠加时使用。自动识别品类→匹配风格→场景感知合成→统一文字→自动质检→多平台适配→标准化交付。E-commerce product image studio: category detection, style matching, scene-aware compositing, text overlay, quality check, multi-platform adaptation, batch delivery.
文献综述 Agent — 解读先行：逐篇固定框架解读 → 用户确认 → A/B 两页结构生成综述 PPT。3 个 skill：paper-survey / paper-interpret / paper-deck。
AI 不缺聪明，缺纪律：交付前自检、跑偏纠偏、系统化调试、多 agent 接力。4 个单文件 SKILL，装上即拷进 ~/.dsh/skills/。
DSH 插件/skill 发布与传播流程——发布前检查、GitHub 建仓推送、索引提交、skill 市场提交（含全部渠道链接）、传播优化（描述/标签/slogan）、发布日志与传播复盘；一次发布铺开多层渠道，持续打造影响力。
Lenovo professional tools for DeepSeek Harness: battery health plus Windows device, performance, storage, Wi-Fi, app and safe-action tools.
Media Forge: agent + skills workflow for media-generation APIs, with a sidebar panel for loading saved generation skills into a conversation
Write the document a newcomer actually needs on day one: what this project is, how to run it, where things live, the five unwritten rules, and who to ask — discoverable, current, and readable in ten minutes, maintained as the cost of doing the work.
Project handoff skill: a dated handoff.md per project — current state, decisions with reasons, pitfalls as symptom/cause/fix, actionable next steps — so any fresh session resumes without archaeology. Bilingual EN/CN.
What will your agent's context cost before it runs? Audit instruction files, skills, MCP tool schemas and memory per source — with duplicates, always-on versus on-demand costs, and a budget you can enforce in CI.
Triage a scattered backlog into a bounded queue of decisions: every item gets a disposition (now / next / scheduled / waiting / reference / drop) with a stated reason, dropped items leave the list instead of haunting it, and anything that changes a commitment goes out as one approval batch.
Pre-send verification for deliverables: opens the real file to verify row and dedup counts, scans for placeholder residue, diffs incremental files against their stated baseline, and logs the batch in a ledger with candidate vs sendable counts kept separate.
Harness-neutral Traditional Chinese enforcer + offline converter (skill, DSH bundle, CLI). One script axis either way + Cantonese/Japanese-only filters. Converts both ways; optional wording preference. Also: an LLM output guard, cp950/GBK console scan.
Complaint-driven competitor research: enumerate competitors, extract features/pricing/positioning plus cited user complaints from reviews, forums, and issue trackers, cluster the complaints, and deliver a comparison table with copy-worthy and skip lists.
DSH plugin: inject a one-line hint naming skills that match the current turn, so installed-but-forgotten methodology skills actually get used. Deterministic matching, no LLM calls, fail-open.
Write changelog entries while the change is in hand: user-facing entries in who/what/how-to-upgrade shape, actions separated from fixes, breaking changes flagged at the top, and a verify pass against the actual diff so the log never claims what the code does not do.
Lightweight decision records for any project: capture decisions that are expensive to reverse or re-litigate — decision, alternatives, the why, reversibility — append-only, supersede instead of rewrite, kept where the project lives so the next session inherits the reasoning, not just the conclusion.