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/affandar/duroxide/copilot-instructionsgit clone --depth 1 https://github.com/affandar/duroxideWhat 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 | $0.01874 | $0.01874 |
| Opus 5 | $0.00937 | $0.00937 |
| Sonnet 5 | $0.00375 | $0.00375 |
| Haiku 4.5 | $0.00187 | $0.00187 |
Grade A, and why
duroxide copilot-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 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 — 169 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Duroxide AI Agent Instructions
A durable execution runtime for Rust. Key docs: docs/ORCHESTRATION-GUIDE.md, docs/provider-implementation-guide.md.
⚠️ CRITICAL: Git Operations
NEVER commit or push without explicit user permission. Always ask before running git commit or git push.
⚠️ CRITICAL: Test Execution
ALWAYS use nextest for running tests. This is mandatory - nextest provides faster execution and better output.
cargo nt # Run all tests (nextest alias)
cargo nt -E 'test(/pattern/)' # Filter tests by pattern
cargo nt --test specific_test_file # Run specific test file
Only fall back to cargo test if nextest is not available. Never use cargo test when cargo nt works.
Architecture Overview
Two-queue message-driven runtime:
- OrchestrationDispatcher (orchestrator queue): Processes workflow turns, durable timers, and external events
- WorkDispatcher (worker queue): Executes activities with automatic lock renewal for long-running work
Core types:
OrchestrationContext- Schedules work viaschedule_activity(),schedule_timer(),schedule_wait(),schedule_sub_orchestration()- all return futures you can.awaitdirectlyEvent/Action- Immutable history entries; providers only store, never generate IDsProvidertrait - Storage abstraction with peek-lock semantics (src/providers/mod.rs)
Critical Patterns
Awaiting schedule_ methods:*
// schedule_* methods return futures that can be awaited directly
let result = ctx.schedule_activity("Task", input).await?;
ctx.schedule_timer(Duration::from_secs(5)).await;
let event_data = ctx.schedule_wait("MyEvent").await;
let sub_result = ctx.schedule_sub_orchestration("Child", input).await?;
Use ctx.join/select, NOT tokio::join/select:
// ✅ CORRECT - deterministic, history-ordered resolution
match ctx.select2(timer, activity).await {
Either2::First(()) => { /* timer won */ }
Either2::Second(result) => { /* activity won */ }
}
let results = ctx.join(vec![f1, f2, f3]).await;
// ❌ WRONG - non-deterministic, breaks replay
let (a, b) = tokio::join!(f1, f2); // NEVER use tokio::join in orchestrations!
tokio::select! { ... } // NEVER use tokio::select in orchestrations!
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 · 169 lines · 1,874 tokens per session scan A 8d5781b11b62
duroxide copilot-instructions.md is an instructions file published in the GitHub repository affandar/duroxide (20 stars, last pushed 6mo ago), licensed MIT. It adds 1,874 tokens to every session, about $0.0094 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
apflow AGENTS.md
AGENTS.md instructions for aiperceivable/apflow, covering project overview, dual model: structure tree + execution dag, five task creation modes, v2 architecture (0.20.0) and key dependencies (embedded, not external layers).
apflow CLAUDE.md
Claude Code instructions for aiperceivable/apflow, covering project overview, dual model: structure tree + execution dag, five task creation modes, v2 architecture (0.20.0) and key dependencies (embedded, not external layers).
apflow copilot-instructions.md
Copilot instructions for aiperceivable/apflow, covering project overview, core principles, python code quality, readability and types (mandatory).
workflowbuilder CLAUDE.md
Instructions for synergycodes/workflowbuilder, covering workflow builder, quick reference, agent signals, monorepo structure and per-workspace docs.
WorkFlowX CLAUDE.md
Instructions for TreeX-X/WorkFlowX, covering claude.md - workflowx instructions, routing, orchestrator workflow, constraints and file operations.
STAR AGENTS.md
AGENTS.md instructions for wanghao9610/STAR, covering agent instructions, 1. think before coding, 2. simplicity first, 3. surgical changes and 4. goal-driven execution.