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/fluttersdk/wind/dynamicgit clone --depth 1 https://github.com/fluttersdk/windWhat 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 | $0.01267 | $0.01267 |
| Opus 5 | $0.00633 | $0.00633 |
| Sonnet 5 | $0.00253 | $0.00253 |
| Haiku 4.5 | $0.00127 | $0.00127 |
Grade A, and why
wind dynamic.instructions.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 yesterday.
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 — 98 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Dynamic JSON renderer (lib/src/dynamic/)
WDynamic renders Flutter widget trees from JSON. Pieces:
WDynamic— the StatefulWidget consumers use. Composes the other parts.WDynamicConfig— security + customization (denyWidgets, builders, customIcons, maxDepth, callbacks).WDynamicState— id-keyed value store + listeners (ChangeNotifier).WDynamicController— wraps external state; tracks_ownsStateto gate disposal.WActionHandler— bridges JSON actions to Dart callbacks; handles_valueinjection.WDynamicRenderer— the dispatch loop; one giant switch overtypestrings.
JSON schema
{
"type": "WText",
"props": {
"text": "Hello",
"className": "text-lg",
"id": "greeting",
"onTap": {"action": "navigate", "args": {"route": "/home"}}
},
"children": [...]
}
type(String, REQUIRED) — must be in the whitelist (defaultWindWidgets∪defaultFlutterWidgets∪config.builders.keys) and not indenyWidgets.props(Map, optional) — widget-specific.className,id, action props, and per-widget config live here.children(List, optional) — nested widget definitions.
Security model (whitelist first)
Order of resolution in WDynamicConfig.isAllowed(type):
config.builders[type]— custom builder always wins.config.denyWidgets.contains(type)— explicit block; emits error widget.- Default whitelist (13 Wind widgets + 16 Flutter core).
Untrusted JSON: pass denyWidgets for surfaces a remote source must not invoke (e.g., 'Container' if you don't trust raw boxes). Always set maxDepth (default 50) — recursion bomb mitigation.
State binding (id-keyed widgets)
WInput, WCheckbox, WSelect, WDatePicker with props.id read initial value from state.get(id) and write back on change. Reactive: same id from another widget triggers a rebuild via WDynamicState's addIdListener.
External vs internal state ownership:
WDynamic(controller: WDynamicController())—_ownsState = false; the host owns disposal.WDynamic(...)no controller —_ownsState = true;WDynamiccreates the state and disposes it.
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.
- yesterday First seen · 98 lines · 1,267 tokens per session scan A 3e91bfaa589e
wind dynamic.instructions.md is an instructions file published in the GitHub repository fluttersdk/wind (33 stars, last pushed 2d ago), licensed MIT. It adds 1,267 tokens to every session, about $0.0063 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
mobile CLAUDE.md
Claude Code instructions for lichess-org/mobile, covering claude.md, project overview, development setup, initial setup and install dependencies.
dart_agent_core AGENTS.md
Instructions for memex-lab/dart_agent_core, covering agents.md, common commands, transcript viewer cli (reads eval traces), two public libraries and architecture (big picture).
sanad-agent AGENTS.md
Instructions for EastStarAI/sanad-agent, covering sanad agent repository contract, 1. documentation hierarchy, rules for ai agents and developers, 1.1. the strict separation pact and 1.2. the living project wiki & incremental documentation.
terradart CLAUDE.md
Instructions for nozomi-koborinai/terradart: Use AGENTS.md as the shared project guide. Read CONTEXT.md for project vocabulary before design work.
dart-sdk-skills AGENTS.md
Instructions for RandalSchwartz/dart-sdk-skills, covering ai agent developer handbook (agents.md), 🎯 repository purpose, 🔄 runbook 1: updating for new dart sdk releases, 🔄 runbook 2: updating for new flutter framework releases and 📏 quality & style guidelines.
dart-expert-skills copilot-instructions.md
Instructions for Poorgramer-Zack/dart-expert-skills, covering flutter skills repository - copilot instructions, repository overview, architecture, directory structure and skill structure.