Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add instructions/criptogus/hermesoffice/claude-mdgit clone --depth 1 https://github.com/criptogus/HermesOfficeWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00627 | $0.00627 |
| Opus 5 | $0.00313 | $0.00313 |
| Sonnet 5 | $0.00125 | $0.00125 |
| Haiku 4.5 | $0.00063 | $0.00063 |
Grade A, and why
HermesOffice CLAUDE.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
This is a copy
100% identical to genoffice CLAUDE.md — 0 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
How it starts
The opening of the file, as written. The whole thing — 48 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
Guidance for AI agents and human contributors working in this repo.
Theming rules (mandatory)
The suite supports light / dark / system UI themes. The switching mechanism is a
data-theme attribute on <html> plus CSS custom properties defined once in
packages/ui/src/tokens.css (light defaults in :root, overrides in
[data-theme='dark'], and a prefers-color-scheme media-query fallback for
system mode).
- UI chrome colors must use semantic tokens. Never write raw
#hex/rgb()in renderer CSS rules or chrome-related inline styles — referencevar(--surface),var(--text),var(--hover), etc. frompackages/ui/src/tokens.css. Raw values are allowed only on custom-property definition lines (--x: #...;— token, accent, or app-scoped variable definitions). CI enforces this for new/changed renderer CSS lines (tools/check-theme-colors.mjs). - Every new token gets both values. Adding a token means adding it to all
three blocks in
tokens.css(light, dark, system-dark fallback). - Accent colors stay per-app. Each app defines
--accent/--accent-dark/--accent-soft(and its dark-adjusted values) in its ownstyles.css. Shared rules referencevar(--accent)and inherit the app's brand color. - Document content never follows the theme. Page surfaces, cell fills, slide content, PDF page bitmaps, export/print stylesheets, chart palettes, highlight color maps, stamps, and WordArt presets are document data: they stay hardcoded, must not reference chrome tokens, and must render/export identically in both themes. (Word-style "dark chrome, white paper".)
- Canvas-drawn UI affordances go through a constants table. Konva/canvas
editing chrome (selection frames, guides, handles) reads from the app's
canvas color table (e.g.
canvas-colors.ts) keyed by the current theme — no inline hex in draw calls.
Build gotchas
- App main-process code (
apps/*/src/main) is compiled into the shell build. After changing it, rebuild the shell or the change silently does not run. - In dev mode, preload changes require a rebuild — a stale preload leaves the renderer blank.
- Workspace packages listed in an app's
dependenciesmust also be added to theexternalizeDepsPluginexcludelist, or the packaged app crashes on launch. useI18n()'stis not referentially stable; never put it in a hook dependency array. Store the key and translate at render time.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 48 lines · 627 tokens per session scan A 595f7e8c825c
HermesOffice CLAUDE.md is an instructions file published in the GitHub repository criptogus/HermesOffice (541 stars, last pushed 19d ago), licensed Apache-2.0. It adds 627 tokens to every session, about $0.0031 per session on Opus 5. A static security scan graded it A with 0 findings. It is 100% identical to genoffice CLAUDE.md, differing in 0 lines, and is treated as a copy.
Other instructions, from other repositories
hermes-webui AGENTS.md
Instructions for nesquena/hermes-webui, covering agent instructions for hermes webui, read first, onboarding and reinstall support, contribution style and before you open a pr — the change guidelines.
genoffice CLAUDE.md
Claude Code instructions for genspark-ai/genoffice, covering claude.md, theming rules (mandatory) and build gotchas.
limboo CLAUDE.md
Instructions for limboo-ai/limboo, covering claude.md, 1. what is limboo?, 2. tech stack (current), 3. project structure and the three electron contexts (critical mental model).
Knote AGENTS.md
Instructions for 1661169091kiwi/Knote, covering agents.md, 构建与验证约定（用户明确要求，所有会话必须遵守） and 测试说明.
AgentCore AGENTS.md
Instructions for Lawofall/AgentCore, covering agents.md, 先读哪里, docs 与 rules and 开发 / 测试（最短）.
duet-agent AGENTS.md
Instructions for dzhng/duet-agent, covering agent guidelines, treat types as documentation, keep names current, keep comments useful and prefer direct, local guarantees.