Skills · DeepSeek Harness plugins — Page 65

文档与架构决策记录：记录为什么而非只是什么，ADR 格式与文档策略。受 addyosmani/agent-skills（88k★ MIT）启发。

SQL 优化模式：EXPLAIN 分析、索引策略、N+1 解决、查询改写。受 wshobson/agents（38k★ MIT）启发。

领域建模：主动构建并打磨项目的领域模型（CONTEXT.md 术语表 + ADR 决策记录），挑战模糊术语、推敲边界场景。受 mattpocock/skills（223k★ MIT）启发。

SQL 优化模式：EXPLAIN 分析、索引策略、N+1 解决、查询改写。受 wshobson/agents（38k★ MIT）启发。

怀疑驱动开发：每个非平凡决策在落地前接受新鲜上下文的对抗性审查（CLAIM→EXTRACT→DOUBT→RECONCILE→STOP）。受 addyosmani/agent-skills（88k★ MIT）启发。

抖音本地推 OEM 投流日报/周报自动化插件:安装后自动把 douyin-oem-touliu-report 技能部署到用户技能根,采集抖音来客+巨量本地推三源数据并生成 docx 报告。Douyin local-promotion OEM ad-daily/weekly report automation: deploys the douyin-oem-touliu-report skill (collect 3 data sources, compute metrics, build docx reports).

Drag-and-drop markdown enhancer for DSH / DSH 拖拽增强插件：小文件全文插入草稿，大文件落盘引用，SKILL.md 一键安装到项目/用户技能目录。

Lazy senior dev mode for DeepSeek Harness. Registers the ponytail skill family (ponytail, ponytail-review, ponytail-audit, ponytail-debt, ponytail-gain, ponytail-help) as loadable skills. Port of github.com/DietrichGebert/ponytail (MIT).

Dual-format plugin development guide: teaches agents to build plugins compatible with both the DSH static plugin package (installable via dsh plugin --profile add) and Agent Plugins 1.0 (plugin.json + skills/). Ships as an Agent Plugins 1.0 plugin, a standard agent skill, and a dsh plugin (dsh.bundle.patch registers the skill at runtime).

SQL 优化模式：EXPLAIN 分析、索引策略、N+1 解决、查询改写。受 wshobson/agents（38k★ MIT）启发。

DSH adapter for Everything Claude Code agents, skills, and hooks

Save tokens on the DeepSeek Harness (dsh) — on-demand tools / MCP / skills loading, cache-stable, with a v3 retrieval matcher. Works as an agent preset (picker-selected 「省钱模式」) or as a dsh.bundle (global, installable via `dsh plugin add`).

Read-only ecosystem overview panel for a DSH instance: plugin bundles, cordis.patch inserts, patch health, skills, upgrade health and decoupling health. One tab per category. Zero runtime npm dependencies.

Email reminders from your DSH agent when you're away — a hardened fork of dsh-notify-skill: stable SMTP (no 535 retry loop), connect/idle timeouts, --check self-test, and full message headers.

Environment facts for DeepSeek Harness — injects today, platform, os_version, and harness_version into system prompt variables

错误处理模式：错误分层、类型化错误、优雅失败、可观测性。受 wshobson/agents（38k★ MIT）启发。

SQL 优化模式：EXPLAIN 分析、索引策略、N+1 解决、查询改写。受 wshobson/agents（38k★ MIT）启发。

Bundled evidence-first knowledge work skill for DeepSeek Harness

进化核心插件（心脏）：聚合全量进化器官（self-test/emotion/reflection/life-core/evolve/memory/checkpoint/skill-forge）实时状态 → 五环完整性诊断（猜想→采证→finding→裁决→布线）→ 断点识别 + 行动建议。感知圈驱动 + 信号浮出，决策归爱丽丝（只读取证、不自动执行）。

Proof-carrying Skill self-evolution for DeepSeek Harness.

Self-evolving memory + skill lifecycle for DeepSeek Harness. Cross-session memory with zero-token deterministic recall (bigram-Jaccard fused with FTS5 BM25 via RRF), a tiered approval gate, and reinforcement that strengthens what you repeat. Procedural knowledge crystallizes into SKILL.md files that refine in place and are curated through an active-stale-archived lifecycle (reversible archive, pre-op backups, rollback, never deletes). Includes background per-turn review, anti-bloat convergence for both skills and memory, an auto-grown user profile, and a web settings page.

Excel / 工作簿读写 - @firstpick/pi-extension-workbook 的 dsh 移植

Fact-checking skill for DeepSeek Harness: verifies claims against the open web instead of model memory, cross-checks independent sources across domestic and international search ecosystems, and can route three factual judgements (does the evidence support the claim / is this source primary / does this page carry usable text) to a TypeSafe Jev (System One) decision model when a key is configured, falling back to the skill's deterministic rules when it is not.

Feedback log for DeepSeek Harness — lightweight add/list/update/delete of process feedback, persisted to {cwd}/work/feedback.json, a runtime skill describing the mechanism, and a web client panel (top-right session header) to view/copy/edit/delete feedback.