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/kokodak/hanji/agents-mdgit clone --depth 1 https://github.com/kokodak/hanjiWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/kokodak/hanji/agents-md)<a href="https://agentmods.dev/instructions/kokodak/hanji/agents-md"><img src="https://agentmods.dev/badge/instructions/kokodak/hanji/agents-md.svg" alt="Measured on agentmods" height="20"></a>What 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 | $0.00960 | $0.00960 |
| Opus 5 | $0.00480 | $0.00480 |
| Sonnet 5 | $0.00192 | $0.00192 |
| Haiku 4.5 | $0.00096 | $0.00096 |
Grade A, and why
hanji AGENTS.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 4d 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.
How it starts
The opening of the file, as written. The whole thing — 72 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
Project Overview
Hanji is a lightweight, local-first Markdown editor built in Rust. It keeps local Markdown files as the source of truth and presents them through a focused visual editing surface.
The editor is designed for simple writing: open a Markdown file, edit it directly, and keep the saved document readable outside Hanji.
Product North Star
Hanji is a light, local-first Markdown editor. It should feel simple enough for notes, durable enough for plain text writing, and open enough for user plugins.
Project Structure
- docs/ is the documentation index and taxonomy.
- docs/architecture/ describes current component ownership and runtime flows.
- docs/design/ defines durable editor concepts and behavior.
- docs/reference/ describes current APIs and supported behavior.
- docs/development/ contains contributor and release workflows.
- docs/plans/ contains proposed or incomplete work.
- docs/decisions/ records durable architectural decisions and rationale.
- site/ contains the static project website deployed with GitHub Pages.
- crates/hanji-core owns text editing primitives: text buffers, selections, transactions, undo, and core commands.
- crates/hanji-markdown owns Markdown parsing, source mapping, projection, and syntax-dependent editing policy.
- crates/hanji-editor owns the portable editor facade used by GPUI and future WebAssembly adapters.
- crates/hanji-storage owns local file and document session behavior.
- crates/hanji-plugin-api is reserved for future public plugin contracts.
- apps/hanji contains the GPUI desktop application.
Engineering Guidelines
- Prefer small, explicit modules over framework-heavy abstractions.
- Keep the default app fast, offline-capable, and understandable.
- Treat plugin APIs as public contracts. Document them before broadening them.
- Avoid introducing persistent services, telemetry, or network behavior without a design document.
- Keep UI copy calm, brief, and useful.
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.
- 4d ago First seen · 72 lines · 960 tokens per session scan A f4550108fcc2
hanji AGENTS.md is an instructions file published in the GitHub repository kokodak/hanji (3 stars, last pushed 1mo ago), licensed MIT. It adds 960 tokens to every session, about $0.0048 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other instructions, from other repositories
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.
speclens CLAUDE.md
Instructions for dansreis/speclens, covering claude.md, what it is, stack decisions, gates and file map.
markra AGENTS.md
AGENTS.md instructions for markrahq/markra, covering markra engineering guidelines, package management, tech stack, code organization and codemirror editor invariants.
markout CLAUDE.md
Instructions for maxmilian/markout, covering claude.md, build, test, run, full test suite (swift testing), a single suite or test and build and launch the app.
mdread CLAUDE.md
Instructions for techjewel/mdread, covering claude.md, what this is, commands, architecture and styles (scss).
MarkCardStudio AGENTS.md
AGENTS.md instructions for pangxiaobin/MarkCardStudio, covering repository guide for coding agents, product scope, stack and commands, architecture and important invariants.