Guix-configs

Diagnose and fix the dual-channel memory problem in Hermes Agent — when `memory.provider: holographic` (or any external provider) is configured but the agent keeps writing only to `memories/MEMORY.md`, bypassing it. Covers why agents default to the markdown file, the four root causes of routing failure, and three remediation options (SOUL.md rules / disable MEMORY.md / mirror via code). Triggers on 'memory 不生效', 'holographic 没在用', 'agent 都写到 MEMORY.md', '双 memory 系统', 'memory tool vs fact_store', 'on_memory_write', 'memory drift', 'Mirrors 不到', or any 'why is my agent ignoring the configured memory provider' question.

Use when the user works inside ~/Projects/Config/Guix-configs and mentions '改 dotfiles', 'blue home', 'guix system reconfigure', 'shepherd service', 'stow 死链', 'blue stow', 'AGENTS.md 翻新', 'fcitx', 'IME 没输入法', 'Electron 没输入法', 'wireplumber', '二轨 dotfiles', '恢复被删除的 dotfiles', or related. Ten sub-protocols — dotfiles deploy verification, worker delegation, multi-line edit safety, Guix service debugging, GNU Stow mutable config, restoring deleted dotfiles, ISO 移植, 需求澄清, 模块归属陷阱.

跨 agent KB 健康度维护。**触发信号**：每周/长会话后例行维护 / 卡片 >50 张 / 检索质量明显下降 / 用户触发 `/agenote-curate` / 发现重复卡片或矛盾结论 / 新增了对话抽取源。**当上述任一信号出现时立即调用本 skill** 做健康检查+去重+归档+权重重分配+reconcile 多源 memory。基础用法见 `agenote-base`；会话中单次经验记录见 `agenote-review`。

Maintain the personal Guix channel at ~/Projects/Config/jeans (Just Enough AI-geNerated Slops). Use when the user asks to "fix the CI build failure issue", "升级 X 包", "add a package", "check upstream updates", "跑 maak upgrade", "修复 auto-update 流水线的 issue

How to author a Hermes Agent skill the right way. Covers the two non-negotiable structural principles every skill must follow — **self-contained** (all runnable artifacts ship inside the skill directory so backup = usable) and **progressive disclosure** (SKILL.md is a thin router; details live under `references/`, `templates/`, `scripts/`) — the directory layout, file-type rules, decision trees for what goes in SKILL.md vs. a support file, **which of the 12 existing categories a new skill belongs to (never top-level `<skill-name>/`)**, and a checklist before declaring a skill done. Trigger when: writing a new skill, refactoring an existing skill's structure, wondering 'should this go in SKILL.md or references()' / 'which category directory should this skill live into', preparing a skill for backup/share, noticing a self-contained violation, or auditing existing skills. Also trigger when the user complains something is 'too verbose' or 'in the wrong place' — those are progressive-disclosure violations.

When the user says '校对文档 / 检查文档是否还有效 / sync doc with code / 扩写 README / 改写文档 / 把这个 PLAN/plan/任务书做成文档 / 文档并入主文档', or when a `refs/*.md` document may be drifting from the tool/CLI/源码 it describes, or when a PLAN file needs to become a user-facing reference doc. Covers three entry points — check, rewrite, plan-to-doc — backed by a `scripts/doc-check.py` checker that catches silent drift between a doc and its source-of-truth (CLI flags, MCP tool names, file paths). Trigger when the user names a doc path and asks whether it's still accurate, when adding a new CLI/MCP/API surface that old docs don't mention, or when a PLAN/任务书/thread needs to graduate into a stable reference.

Stress-test a plan or design doc by walking the decision tree one branch at a time with the user. Triggers when the user says "压力测试这个方案 / 把这个计划拍一遍 / 我有个 plan 想跟你对一下 / 等等我有问题要问你 / grill this / 拍砖 / 我总觉得哪里不对 / 这个方案经不经得起推敲" or any other "challenge my existing plan" framing. Distinct from `task-planner` (which builds a plan from scratch) and `architecture-advisor` (which gives a second-opinion read on a finished design). The output is a numbered decision ledger (D1, D2, ...) that either becomes the new plan or marks which sections of the existing plan need rewriting.

调试用户的 Emacs Lisp 配置 —— 键位/face/模式行为不符合预期时使用。覆盖 keymap 优先级链、describe-key / C-h k 输出解读、emacsclient --eval 实测运行 daemon、定位"包内置绑定/主题覆盖"等典型坑。触发：「按键不生效」「C-x 在 X mode 行为不对」「我绑了 Y 但按下去是 Z」「global-set-key 没生效」「face 被主题改回去」「reload 不生效」。

Weekly red-light audit of Hermes Agent's own config. Triggers on agent slow / dumber over time, duplicate injection, zombie skills, state drift, monitor lying.

Weekly 14-check red/green audit of agent config bloat. Diagnoses inject size, zombie skills, state drift, monitor honesty, log bloat, secret leaks.

Force a post-edit walkthrough of every file changed, line by line.

Use when the user wants to add, remove, prune, reorganize, audit, or curate the Hermes Agent skill library on disk. Triggers: '精简 skill', '删掉 XX skill', 'skill 太多', '哪些 skill 在用', '清理 Hermes skill', 'skill 审计', 'trash skill 目录', 'bundled skill 删不掉', 'curator 暂停', 'archiving skills', '分类 skill', '重新分类', 'skill 太散', 'skill 没分类', '重组 skill', 'reorganize skills', '创建分类目录', '新建分类', or any bulk add/remove/regroup of skill directories. Covers: skills_list 视图 vs 磁盘真实差异、bundled read-only 0555 权限位破解、XDG trash 范式(用户偏好 rm→trash)、用户 dotfile 源(Guix stow)与 Hermes 安装源(bundle)边界、分类重组(伪分类→真分类)、git 索引状态对 mv 的影响、作用域边界(本 skill 默认只管 ~/.local/share/hermes/skills/,不动 ~/.config/agents/skills/ 或 ~/.hermes/skills/ 除非用户明确授权)。Companion to skill-authoring: this skill handles *reorganizing existing* skills, while skill-authoring §9 routes placement for *new* skills.

"Repair or assemble the audio+subtitle track for an already-rendered, segmented video (HyperFrames explainer, slide deck, course video, screencast) so narration matches the picture. Trigger on 音频和视频对不上, 配音和画面不匹配, narration drift, voiceover out of sync, 音画不同步, 重做配音, 重新对位, 补字幕, 烧字幕. Covers topic drift, length drift, boundary drift. Pipeline: 1fps probe → visual inventory → script rewrite → TTS → atempo/padding → silencedetect SRT → ffmpeg subtitles burn-in.

Use when the user starts a new project, faces ambiguous requirements, says 'I don't know what I don't know', wants discovery/research before design. Triggers: '识别未知', '设计前的调研', '验证假设', '用户访谈设计', 'discovery phase', 'validate assumptions'. Distilled from Than Tibbetts' 'A Field Guide to Fable: Finding Your Unknowns' — gives a 5-stage workflow for finding unknowns before you can find their solutions.

为外部 agent 工具的本地会话库(zcode SQLite / Claude Code JSONL / opencode SQLite / Codex sessions / Crush / pi-agent 等)做 schema 侦察,产出 importer 需要的字段映射、JSON 样本、时间单位、system-reminder 折叠规则、子会话嵌入策略。**触发信号**:用户说"做 importer / 把 X 的会话导进 hermes / 摸清这个 DB 的 schema / 从 X 拉历史记录 / 把 zcode 历史迁过来"等。属于"输出可消费的 schema 情报",不是 source-code 侦察(codebase-scout)也不是 prompt 迁移(importing-agent-prompts)。

Extract crush CLI sessions from per-project SQLite DBs.

子任务派发规范 —— 定义 hermes 主会话如何把任务拆给 delegate_task 子 agent,以及子 agent 应如何产出标准化的 handoff 报告。适合在需要把大任务并行化、或需要独立验证的子任务时加载使用。

严格的代码审查助手 —— 对代码 diff、文件、PR 或整个模块做基于证据的对抗性审查,覆盖架构、代码质量、工程实践、性能与安全四维。当用户说"review/审查/检查这段代码"、"看看我写的对不对"、"code review"、"提 PR 前帮我看看"、"这有什么问题"时触发。

快速只读代码侦察 —— 扫描仓库定位文件、模式、依赖关系,输出压缩的结构化发现供后续工作消费。当用户说"扫一下这个仓库 / 找一下 X 在哪 / 摸清项目结构 / 摸清这个模块怎么工作的 / 这段代码涉及哪些文件"时触发。

用 HyperFrames 制作 Emacs Lisp 教学视频的完整工作流——从文档提取概念→设计可视化(画数据结构+算法流程)→编写 HTML/SVG/GSAP 时间线→渲染 MP4。

架构顾问 —— 质疑假设、评估方向、提供替代方案与架构模式建议,做"第二意见"。当用户说"这个设计合不合理 / 帮我看看架构方向 / 有没有更好的方案 / 在错误方向投入太多之前拉一下 / 评估下技术选型"时触发。

战略规划师 —— 面试式澄清需求、评估架构可行性、对比替代方案、产出可执行的实施计划(任务清单 + 依赖图 + 风险评估)。当用户说"帮我规划一下 / 拆解任务 / 制定方案 / 怎么实施 / 制定计划 / 任务分解"时触发。

文档检索专家 —— 收集外部信息、API 文档、最佳实践、版本兼容性数据,为技术决策提供信息支撑。当用户说"调研一下这个库 / 查 API 文档 / 收集最佳实践 / 看看这个库怎么用 / 版本兼容性"时触发。

Lint repo docs for single-source-of-truth violations.

跨 agent 共享知识库（agenote）—— 任务开始时查、过程中复用、结束时记。**触发信号**：开始非平凡任务前 / 遇到已踩过或疑似踩过的坑 / 联网查到新方案 / 用户纠正/纠正自己 / 长任务结束。**当上述任一信号出现时立即调用本 skill**，按其内部规则决策（list→search→get；add→touch；commit）。需要维护 KB 健康度（去重/归档/重整）或从 6+ agent 抽取对话 reconcile 时转 `agenote-curator`；会话结束需评估可记录经验时转 `agenote-review`。底层 CLI 为 `agenote`（`~/.local/bin/agenote`），与 MCP tool 行为对齐。

会话后经验采集与留痕。**触发信号**：agenote-hooks 插件检测到完成信号 / 用户触发 `/agenote-summarize` / 长任务结束前的例行评估 / 用户纠正了你 / 排查 >2 步才定位根因 / 发现了比初始方案更优的写法。**当上述任一信号出现时立即调用本 skill** 做"经验信号识别 + ENTRY_TYPE 判定 + 留痕决策树"，按其内部规则写卡片或 touch 已有卡片。基础用法见 `agenote-base`。

Import conversation history from a third-party coding-agent CLI/TUI (pi, Claude Code, Codex, OpenCode, Aider, zcode, …) into Hermes's local session store so the conversations appear in `hermes sessions list` and can be resumed with `hermes --resume <id>`. Triggers: 'import pi/zcode sessions', '把 claude code / zcode 对话导入 hermes', 'migrate my chat history', '把 .codex/sessions 灌进 hermes', '恢复被删的 dotfiles'. Also handles recap work when hermes's `session_search` is empty but another agent left local state behind. **Before writing any importer against an agent's data dir, run `find ~ -name '*.db' -o -name '*.sqlite*' 2>/dev/null | grep -i <agent>` first — many agents (zcode, pi context-mode) keep the real session tree in a SQLite DB even when an obvious `*.jsonl` directory exists nearby.** For the related but narrower task of generating a schema reconnaissance report without writing the importer, see `agent-history-importer`.

Typeset professional documents and product landing pages: resumes, one-pagers, white papers, letters, portfolios, slide decks, landing pages. Warm parchment, ink-blue accent, serif-led hierarchy. CN uses TsangerJinKai02, EN uses Charter, JA uses YuMincho (best-effort). Triggers on "做 PDF / 排版 / 一页纸 / 白皮书 / 作品集 / 简历 / PPT / slides / Marp / markdown slides / マークダウンのスライド / 落地页 / 官网 / landing page / product page", or "build me a resume / make a one-pager / design a slide deck / turn this into a PDF / make this presentable / create a landing page".

Design one-off HTML artifacts (landing, deck, prototype).

Diagnose and fix input method (fcitx5) issues in Electron/Chromium apps running under Wayland compositors (niri, sway, Hyprland). Covers XDG desktop override pattern, Electron flag differences across versions, and fcitx5 diagnostic workflow.

跨系统借鉴设计 —— 当用户说"用 X 系统的成熟设计改 Y / 借鉴上游参考实现 / 把 Y 的 Z 机制搬过来 / 跨 agent 升级 / 上游设计 port 到本机"时触发。覆盖"先深度对位(现状摸底 × 上游 reference × 缺口矩阵)→ 提炼设计原则(每条标上游 file:line)→ 产出 1000+ 行 PLAN.md(分阶段 + schema 草稿 + 验收清单 + 协同点)"这个 class 的工作。

Diagnose TCP services that span multiple hops — rootlessport → container → iptables DNAT → QEMU user networking → guest VM, or any chained proxy / NAT / userspace forwarder. Triggers when a 'TCP service not reachable' symptom has multiple plausible layer-3/4/7 explanations, when `nc -zv` says succeeded but the application still errors, when packet counters on iptables / NAT rules stay at 0, or when the user reports an RDP/VNC/SSH/DB connection failing despite some TCP handshake seemingly working. Covers hop-by-hop verification, what `nc -zv succeeded` actually proves, packet-counter-driven root-causing, and how to spot a host-side listener (rootlessport, socat, haproxy) that has nothing behind it.

辅助中国高校学生完成课程实践报告、专题调研、记录册、上汇报讲稿等 课程作业任务。覆盖"形势与政策""思政课""专业实习"等典型场景。 触发：用户给出课程材料（.docx/.pdf/资料）并要求"根据这份文档写报告/填记录册/ 做讲稿"，或"完成附件1的实践手册"。

把外部 agent 的 system prompt 库(pi agent、Claude Code subagent、Codex/OpenCode/Aider 等的 .md 设定)搬进 Hermes 的 skill 系统。当用户提到 "把这些 agent 设定迁过来 / pi 的 agent 能用在 hermes 上吗 / 把 XX agent 的 prompt 搬过来 / 我有一堆其他 agent 的 markdown 配置 / 我不再用 XX 了但这些 prompt 有价值" 时触发。也涵盖 "提示词模板"、"review 提示词"、"把这段 prompt 注入到会话" 这类需求。

Add captions to a talking-head video. ONE catalog (CATALOG.md) of 32 visual identities behind two engines: column-flow (captions composited INTO the scene — matte occlusion + mix-blend; cream/ink/editorial/keynote/documentary/loud/neon/glitch/chrome/velocity) and themed constitutions (anchor/ordnance/terminal/neonsign/stardust/stomp/scoreboard/transit/vhs/arcade/dossier/laser/thunder/hologram/biolume/aurora/spectrum/papercut/popup/chalkboard/graffiti/brush/inkwater/ransom/lastpage/nightcity — e.g. a glyph-decode climax, a neon sign WRITTEN stroke by stroke, or the quiet `anchor` rail default). Route by identity, never by mode. Trigger on "captions/subtitles", "embed/cinematic captions", "VFX captions", "炸/特效/酷炫字幕", a named identity, or top-tier motion-graphics asks. Embedding every word is wrong for most talking-head content — `anchor` is the verbatim default. Pipeline: transcription → hyperframes remove-background matting → HTML render → ffmpeg overlay. Requires hyperframes and a single-subject clip.

faceless-explainer video workflow - arbitrary text (article / notes / topic / brief) -> narrator_scripts.json + audio (voice + BGM) + section_plan.md -> typography / abstract-graphics / diagram / data-viz video. Typical length up to ~3 min (sweet spot ~30-90s); a genuinely longer piece is general-video, not this workflow. Generates its OWN narration (TTS) — it does not sync to a user-supplied / pre-recorded voiceover (that is general-video). No website capture, no real product screenshots. If the text names a product / its site to promote, that is /product-launch-video; when product-vs-topic is unclear, start at /hyperframes-read-first.

The fallback workflow for authoring custom HyperFrames video compositions at any length or format — longer or multi-scene pieces, brand / sizzle reels, montages, title cards, static loops, and freeform compositions. Input- and length-agnostic. If a specialized workflow clearly fits the input — a marketed product, a website, a topic explainer, a GitHub PR, existing footage, a short motion graphic, or a Remotion port — prefer it (see /hyperframes); use this only as the general fallback when none fit.

All animation knowledge for HyperFrames — atomic motion rules, multi-phase scene blueprints, scene transitions, broader motion-design techniques, AND the seven runtime adapters (GSAP default, plus Lottie, Three.js, Anime.js, CSS keyframes, Web Animations API, TypeGPU). Use for any motion or animation task: pick 2-4 rules and compose, or load a blueprint, or look up runtime-specific API (e.g. GSAP eases / Lottie player / Three.js mixer). HyperFrames-native: single paused timeline, seek-safe, deterministic.

HyperFrames CLI dev loop — `npx hyperframes` for scaffolding (init), validation (lint, inspect), preview, render, and environment troubleshooting (doctor, browser, info, upgrade). Use when running any of these commands or troubleshooting the HyperFrames build/render environment. For asset preprocessing commands (`tts`, `transcribe`, `remove-background`), invoke the `hyperframes-media` skill instead.

Non-animation creative direction for HyperFrames videos. Use for design spec (frame.md / design.md) handling, palettes, typography, narration, beat planning, audio-reactive visuals, composition patterns, and brand / style decisions. For atomic motion patterns and scene blueprints, use `hyperframes-animation`.