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/channprj/kmsg/agents-mdgit clone --depth 1 https://github.com/channprj/kmsgWhat 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 | $0.00579 | $0.00579 |
| Opus 5 | $0.00290 | $0.00290 |
| Sonnet 5 | $0.00116 | $0.00116 |
| Haiku 4.5 | $0.00058 | $0.00058 |
Grade A, and why
kmsg 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 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.
How it starts
The opening of the file, as written. The whole thing — 73 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
Build Commands
swift build # Debug build
swift build -c release # Release build
.build/debug/kmsg [command] # Run debug build
.build/release/kmsg [command] # Run release build
Testing
No automated test suite. Test manually:
.build/debug/kmsg status --verbose # Check accessibility & app status
.build/debug/kmsg inspect --depth 5 # Inspect UI hierarchy
.build/debug/kmsg chats --verbose --limit 20 # List chats
.build/debug/kmsg send "Name" "msg" --dry-run # Test send without sending
.build/debug/kmsg read "Chat" --limit 50 # Read messages
Architecture
CLI Commands (Commands/*.swift)
↓
KakaoTalkApp (High-level business logic)
↓
UIElement (Accessibility API wrapper)
↓
macOS Accessibility APIs (AXUIElement)
Key components:
-
UIElement (
Accessibility/UIElement.swift): Wrapper aroundAXUIElementwith breadth-first search, attribute access, hierarchy traversal. Supportslimitparameter for early termination. -
KakaoTalkApp (
KakaoTalk/KakaoTalkApp.swift): Process lifecycle, window management, auto-launch with timeout. Thread-safe viaSendable. -
Commands: Each implements
ParsableCommand. Handle CLI interaction, permission checks, delegate business logic to KakaoTalkApp.
KakaoTalk UI Structure
Main Window (id: Main Window):
├── Navigation: id: friends, id: chatrooms, id: more
├── Unread count: AXStaticText
└── Content: AXScrollArea > AXTable > AXRow > AXCell
Chat Window (title = chat name):
├── Messages: AXScrollArea > AXTable > AXRow
└── Input: AXTextArea or AXTextField
Technical Notes
- Bundle ID:
com.kakao.KakaoTalkMac - Requires Accessibility permission (System Settings > Privacy & Security > Accessibility)
- Keyboard input uses
CGEventfor Korean text (Unicode support) - Search methods support
limitfor early termination - use to avoid full UI traversal - All classes interacting with accessibility APIs should be
Sendable - For version bumps, release tags, or release automation changes, follow
VERSIONING.md
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 · 73 lines · 579 tokens per session scan A e5fc54cab03f
kmsg AGENTS.md is an instructions file published in the GitHub repository channprj/kmsg (256 stars, last pushed 9d ago), licensed MIT. It adds 579 tokens to every session, about $0.0029 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 2026-08-30.
Other instructions, from other repositories
kakaocli AGENTS.md
Instructions for silver-flight-group/kakaocli, covering agents.md — kakaocli for ai agents, prerequisites, first-time setup, store kakaotalk credentials (saved in macos keychain) and verify everything is working.
kakaocli CLAUDE.md
Instructions for silver-flight-group/kakaocli, covering claude.md, build & test, project structure, key technical details and database.
starwind-ui AGENTS.md
Instructions for starwind-ui/starwind-ui, covering agents.md, project shape, useful commands and contribution conventions.
Codex-Skills AGENTS.md
Instructions for TheGoat395/Codex-Skills, covering repository operating notes, required checks, evidence rules and change hygiene.
ACCESSIBILITY.md AGENTS.md
Instructions for mgifford/ACCESSIBILITY.md, covering repository instructions for ai coding agents, scope, when each instruction file is useful, project mission and status and required project context.
monochrome-ui AGENTS.md
Instructions for vaneenige/monochrome-ui, covering agents.md, north stars (non-negotiables), why the core looks weird (and should stay weird), clever tricks and code style.