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/jenreh/appkit/agents-mdgit clone --depth 1 https://github.com/jenreh/appkitWhat 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 | $0.01621 | $0.01621 |
| Opus 5 | $0.00811 | $0.00811 |
| Sonnet 5 | $0.00324 | $0.00324 |
| Haiku 4.5 | $0.00162 | $0.00162 |
Grade A, and why
appkit 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.
This is a copy
86% identical to harmonyhub-py AGENTS.md — 35 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 — 180 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AppKit — Reflex-Mantine Component Library
Stacks: Python 3.14 · Reflex.dev (UI) · FastAPI · SQLAlchemy 2.0 · Alembic · Pydantic · FastMCP · LangChain
1) Golden Rules
- Think → Memory → Tools → Code → Memory. Use code-reasoning; search Memory and claude-context first; minimal diff; write learnings back.
- Tests are truth. Failures → fix code first. Change tests only if clearly wrong spec.
- Minimal diff. Add tests before code. Keep simple.
- Consistency > cleverness. Follow SOPs and stack idioms.
- Memory multiplies. Persist decisions, patterns, error signatures, proven fixes.
- Files ≤ 1000 lines. Exceed → refactor (see §5).
- No extensive docs/summaries/comments unless requested.
- No
--autogeneratefor Alembic migrations; write manually. - No
catto create files; use tools. - Log default:
logger.debug. Important events:logger.info. Issues only:logger.warning/error. Noprint. - Caveman skill applies to all writes here.
Prefer local changes over cross-module refactors.
2) Task Bootstrap Pattern
<!-- plan:start
goal: <one line clear goal>
constraints:
- Python 3.14; Reflex UI; FastAPI; SQLAlchemy 2.0; Alembic; Pydantic; appkit_mantine;
- logging: no f-strings in logger calls
- files ≤ 1000 lines; apply design patterns where appropriate
- minimal diff; add/adjust tests first
definition_of_done:
- tests pass; coverage ≥ 80% (non-Reflex classes & Reflex states); lint/type checks clean; memory updated
steps:
1) Search Memory for "<keywords>"
2) Draft/adjust failing test to capture expected behavior
3) Implement minimal code change
4) Run task test; iterate until green
5) Update Memory: decisions, patterns, error→fix
plan:end -->
3) Tooling Decision Matrix
| Situation | Primary | Secondary | Store to Memory |
|---|---|---|---|
| API/pattern uncertainty | Context7 | — | Canonical snippet + link; edge cases |
| Ecosystem bug/issue | DuckDuckGo | Context7 | Minimal repro; versions; workaround |
| Repeated test failure | Memory (search) | Context7 | Error signature → fix; root cause |
| New feature scaffold | Context7 | — | How‑to snippet; checklist |
| House style/tooling | This file | Context7 | Checklist results |
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 · 180 lines · 1,621 tokens per session scan A e0c973c514a5
appkit AGENTS.md is an instructions file published in the GitHub repository jenreh/appkit (4 stars, last pushed 6d ago), licensed MIT. It adds 1,621 tokens to every session, about $0.0081 per session on Opus 5. A static security scan graded it A with 0 findings. It is 86% identical to harmonyhub-py AGENTS.md, differing in 35 lines, and is treated as a copy.
Other instructions, from other repositories
codex-image-context-runtime AGENTS.md
Instructions for shixinnt/codex-image-context-runtime, covering agents.md, public boundary, runtime contract and changes.
Omni-Rewriter AGENTS.md
Instructions for WayneJin0918/Omni-Rewriter: Maintainer and coding-agent conventions live in docs/AGENTS.md. Also see .cursor/skills/omni-rewriter-h3-pe/SKILL.md, .cursor/skills/omni-rewriter-model-contribution/SKILL.md, and .cursor/skills/omni-rewriter-promo-pipeline/SKILL.md.
openai-compatible-imagegen AGENTS.md
Instructions for Syh1906/openai-compatible-imagegen, covering agent instructions, project scope, package management, checks and module boundaries.
diffusers AGENTS.md
AGENTS.md instructions for huggingface/diffusers, a project described as: 🤗 Diffusers: State-of-the-art diffusion models for image, video, and audio generation in PyTorch.
comfyui_mcp AGENTS.md
Instructions for hybridindie/comfyui_mcp, covering agents.md, independent & harness-agnostic, project overview, tech stack and project structure.
OGAD AGENTS.md
AGENTS.md instructions for off-grid-ai/OGAD, covering off grid ai desktop — agent guide, design — desktop-first, off grid brand, use the screen real estate — desktop density rules, what this app is and stack.