agentleFS
MovGP0MovGP0/Skills/m07-concurrency/SKILL.md

m07-concurrency

Skill1 starsChanged 4 months agoMITView on GitHub

CRITICAL: Use for concurrency/async. Triggers: E0277 Send Sync, cannot be sent between threads, thread, spawn, channel, mpsc, Mutex, RwLock, Atomic, async, await, Future, tokio, deadlock, race condition, 并发, 线程, 异步, 死锁

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 3 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Before choosing concurrency primitives:

    line 15
  2. Distinctive:

    No sharing → message passing (channels)

    line 43
  3. Distinctive:

    CRITICAL: Don't just fix the error. Trace UP to find domain constraints.

    line 56

≈ “Common” is an estimate: at least 1 in 5 of the 308,232 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Concurrency

Layer 1: Language Mechanics

Core Question

Is this CPU-bound or I/O-bound, and what's the sharing model?

Before choosing concurrency primitives:

  • What's the workload type?
  • What data needs to be shared?
  • What's the thread safety requirement?

Error → Design Question

ErrorDon't Just SayAsk Instead
E0277 Send"Add Send bound"Should this type cross threads?
E0277 Sync"Wrap in Mutex"Is shared access really needed?
Future not Send"Use spawn_local"Is async the right choice?
Deadlock"Reorder locks"Is the locking design correct?

Thinking Prompt

Before adding concurrency:

  1. What's the workload?
    • CPU-bound → threads (std::thread, rayon)
    • I/O-bound → async (tokio, async-std)
    • Mixed → hybrid approach
  2. What's the sharing model?
    • No sharing → message passing (channels)
    • Immutable sharing → Arc<T>
    • Mutable sharing → Arc<Mutex<T>> or Arc<RwLock<T>>
  3. What are the Send/Sync requirements?
    • Cross-thread ownership → Send
    • Cross-thread references → Sync
    • Single-thread async → spawn_local

Trace Up ↑ (MANDATORY)

CRITICAL: Don't just fix the error. Trace UP to find domain constraints.

Domain Detection Table

Context KeywordsLoad Domain SkillKey Constraint
Web API, HTTP, axum, actix, handlerdomain-webHandlers run on any thread
交易, 支付, trading, paymentdomain-fintechAudit + thread safety
gRPC, kubernetes, microservicedomain-cloud-nativeDistributed tracing
CLI, terminal, clapdomain-cliUsually single-thread OK

Example: Web API + Rc Error

"Rc cannot be sent between threads" in Web API context
    ↑ DETECT: "Web API" → Load domain-web
    ↑ FIND: domain-web says "Shared state must be thread-safe"
    ↑ FIND: domain-web says "Rc in state" is Common Mistake
    ↓ DESIGN: Use Arc<T> with State extractor
    ↓ IMPL: axum::extract::State<Arc<AppConfig>>

Generic Trace

"Send not satisfied for my type"
    ↑ Ask: What domain is this? Load domain-* skill
    ↑ Ask: Does this type need to cross thread boundaries?
    ↑ Check: m09-domain (is the data model correct?)
SituationTrace ToQuestion
Send/Sync in Webdomain-webWhat's the state management pattern?
Send/Sync in CLIdomain-cliIs multi-thread really needed?
Mutex vs channelsm09-domainShared state or message passing?
Async vs threadsm10-performanceWhat's the workload profile?

Trace Down ↓

From design to implementation:

"Need parallelism for CPU work"
    ↓ Use: std::thread or rayon

"Need concurrency for I/O"
    ↓ Use: async/await with tokio

"Need to share immutable data across threads"
    ↓ Use: Arc<T>

"Need to share mutable data across threads"
    ↓ Use: Arc<Mutex<T>> or Arc<RwLock<T>>
    ↓ Or: channels for message passing

"Need simple atomic operations"
    ↓ Use: AtomicBool, AtomicUsize, etc.

Send/Sync Markers

MarkerMeaningExample
SendCan transfer ownership between threadsMost types
SyncCan share references between threadsArc<T>
!SendMust stay on one threadRc<T>
!SyncNo shared refs across threadsRefCell<T>

Quick Reference

PatternThread-SafeBlockingUse When
std::threadYesYesCPU-bound parallelism
async/awaitYesNoI/O-bound concurrency
Mutex<T>YesYesShared mutable state
RwLock<T>YesYesRead-heavy shared state
mpsc::channelYesOptionalMessage passing
Arc<Mutex<T>>YesYesShared mutable across threads

Decision Flowchart

What type of work?
├─ CPU-bound → std::thread or rayon
├─ I/O-bound → async/await
└─ Mixed → hybrid (spawn_blocking)

Need to share data?
├─ No → message passing (channels)
├─ Immutable → Arc<T>
└─ Mutable →
   ├─ Read-heavy → Arc<RwLock<T>>
   └─ Write-heavy → Arc<Mutex<T>>
   └─ Simple counter → AtomicUsize

Async context?
├─ Type is Send → tokio::spawn
├─ Type is !Send → spawn_local
└─ Blocking code → spawn_blocking

Common Errors

ErrorCauseFix
E0277 Send not satisfiedNon-Send in asyncUse Arc or spawn_local
E0277 Sync not satisfiedNon-Sync sharedWrap with Mutex
DeadlockLock orderingConsistent lock order
future is not SendNon-Send across awaitDrop before await
MutexGuard across awaitGuard held during suspendScope guard properly

Anti-Patterns

Anti-PatternWhy BadBetter
Arc<Mutex<T>> everywhereContention, complexityMessage passing
thread::sleep in asyncBlocks executortokio::time::sleep
Holding locks across awaitBlocks other tasksScope locks tightly
Ignoring deadlock riskHard to debugLock ordering, try_lock

Async-Specific Patterns

Avoid MutexGuard Across Await

// Bad: guard held across await
let guard = mutex.lock().await;
do_async().await;  // guard still held!

// Good: scope the lock
{
    let guard = mutex.lock().await;
    // use guard
}  // guard dropped
do_async().await;

Non-Send Types in Async

// Rc is !Send, can't cross await in spawned task
// Option 1: use Arc instead
// Option 2: use spawn_local (single-thread runtime)
// Option 3: ensure Rc is dropped before .await

Related Skills

WhenSee
Smart pointer choicem02-resource
Interior mutabilitym03-mutability
Performance tuningm10-performance
Domain concurrency needsdomain-*

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.