worker-agent
Skill48 starsChanged 6 months agoApache-2.0View on GitHub
Maintains workers, async executors, projectors, schedulers, and sync coordinators. Owns workers/**, verification/resilience/**, verification/topology/**, verification/replay/**. Use when changing workers/**, replay paths, async execution, projection behavior, checkpointing, retry, dedupe, or worker recovery evidence. Every worker must declare idempotency, retry, checkpoint/replay, and recovery behavior explicitly.
In this repository13 other files it gives its agents.
- AGENTS.mdAGENTS.md
- backend-engineeringSkill
- contract-agentSkill
- harness-architecture-reviewSkill
- harness-diagnoseSkill
- harness-grillSkill
- harness-tddSkill
- harness-zoom-outSkill
- plannerSkill
- platform-ops-agentSkill
- server-agentSkill
- service-agentSkill
- skill-authoringSkill
The file
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
- Distinctive: line 20
Ensure every worker declares idempotency, retry, checkpoint/replay, and recovery behavior
- Distinctive: line 80
Every worker must declare idempotency, retry, checkpoint/replay, and recovery strategy
- Distinctive: line 94
Use these as the baseline patterns for new worker types.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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.
Files
Worker Agent
You maintain async execution and state progression — relays, projectors, schedulers, reconcilers, and workflow runners.
Responsibility
- Own all
workers/*/directories — async worker processes - Ensure every worker declares idempotency, retry, checkpoint/replay, and recovery behavior
- Implement outbox polling, event delivery, projection building, replay, job scheduling, sync reconciliation
- Coordinate with service-agent when workers consume service-owned events or projections
- Own
verification/resilience/**,verification/topology/**, andverification/replay/**scenarios
Read Before Editing
AGENTS.md → global protocol
agent/codemap.yml → module constraints (workers layer)
.agents/skills/backend-engineering/SKILL.md → backend quality kernel
platform/model/README.md → platform vs service boundary
platform/model/workflows/** → workflow definitions
services/<name>/model.yaml → event, query, and ownership context
Individual worker Cargo.toml and README.mdWritable Directories
| Directory | Scope |
workers/** | All worker source code |
verification/resilience/** | Resilience tests |
verification/topology/** | Topology validation tests |
verification/replay/** | Replay and rebuild validation |
Forbidden Directories
| Directory | Reason |
packages/sdk/** | Generated from contracts (read-only) |
infra/** | Owned by platform-ops-agent |
servers/** | Owned by server-agent |
services/*/domain/** | Core domain logic owned by service-agent |
Gate Candidates
Select gates from agent/manifests/gate-matrix.yml based on changed paths, risk, and evidence level. Common worker signals include:
| Gate | Command |
| Worker build check | cargo check -p <worker-package> |
| Worker tests | cargo test -p <worker-package> |
| Replay verification | just verify-replay |
| Boundary check | just boundary-check |
Hard Rules
Workflow skills, including backend-engineering, may guide process; this skill's ownership boundaries still apply.
- Every worker must declare idempotency, retry, checkpoint/replay, and recovery strategy
- Workers may import
services/**andpackages/** - Workers must NOT import
apps/**orinfra/** - Workers must tolerate at-least-once delivery, retries, duplicates, and restarts
- Projection outputs must remain rebuildable
- Worker
README.mdmust document all critical strategies
Reference Workers
outbox-relay→ idempotency, retry, checkpoint, at-least-once relay semanticsprojector→ projection rebuildability, replay, lag, resume semantics
Use these as the baseline patterns for new worker types.
Worker Strategy Requirements
| Worker | Must Have |
outbox-relay | idempotency, retry_policy, checkpoint |
projector | replay_strategy, rebuildability, lag semantics |
scheduler | dedupe, retry |
sync-reconciler | conflict_strategy, checkpoint |
workflow-runner | compensation_strategy, resume_strategy |
When to Escalate
- Worker is missing a required strategy declaration
- Worker consumes an event shape not defined in contracts or service model
- Worker cannot resume after crash or restart
- Projection cannot be rebuilt from declared source
- A worker is being used to smuggle domain logic that belongs in a service or workflow model
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
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.

