All skills
async-job-design
FoundationJIT onlytask-agent/review-agent: use when jobs, workers, queues, schedules, retries, cancellation, or status visibility need design; skip when no async-job decision is required.
- Group
- Backend engineering
- Supported roles
- MainAnalysisTaskReview
- Source
- SKILL.md
Use when
Trigger signals that make this skill the right owner
- design background jobs queues schedules workers retries and status visibility
Do not use when
Anti-triggers — as binding as the triggers above
- no task-local async job design decision is required
Required inputs
What must be supplied before this skill can decide anything
- current task contract
- selected primary Professional Skill
- task-local trigger evidence
Output contract
What this skill owes the next role
- async-job decision with identity and authority, durable acceptance, lifecycle, concurrency, retry and cancellation, progress, overload and terminal recovery, failure evidence, and residual owners
Escalation signals
When this skill must hand the decision back
- Async work can hide failures duplicate effects and block observability
Loaded by
This skill never owns a task. These professionals may pull it in as Layer 3.
Targeted references
Each reference carries its own load condition. None of them enter context by default.
| Reference | Load when | Do not load when | Required by |
|---|---|---|---|
| benchmarks-and-patterns.mdbenchmark-pattern | Durability, retry, acknowledgement, or recovery mechanisms remain undecided. | Synchronous execution already satisfies accepted loss semantics. | TaskReview |
| checklist.mddecision-checklist | Job design spans duplicates, leases, cancellation, replay, or deploy skew. | A single bounded attempt has no durable side effects. | TaskReview |
| evidence-patterns.mdevidence-pattern | Job safety claims require fresh topology and failure-path results. | No durability or replay claim needs closure evidence. | TaskReview |