agentsclimarketplace

Event sourced orchestration with decider projector

Skill kjuhwa/skills-hub/skills/architecture/event-sourced-orchestration-with-decider-projector

Self-correcting knowledge corpus for Claude Code — 9 stable shape clusters, bias-correction pipeline baked into contribution flow. 47 papers, 45 techniques, 1.1k skills.

Install
npx -y skills add kjuhwa/skills-hub --skill event-sourced-orchestration-with-decider-projector

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Use event sourcing for domain state with pure decider (commands->events) and projector (events->read model) functions

SKILL.md

2.7 KB, 479 tokens by cl100k_base, as published. Nobody here has run it

When to Apply

  • You need an audit trail of all state changes (threads, turns, checkpoints, approvals)
  • You want to replay history without re-running external agents
  • You need to support concurrent requests and rollback failed operations
  • You want domain logic (commands + events) separate from read model updates

Steps

  1. Define command schemas: Thread.Create, Thread.Turn.Start, Thread.Checkpoint.Revert (in contracts)
  2. Define event schemas: ThreadCreated, MessageSent, TurnDiffCompleted (in contracts)
  3. Implement decider function:
    • Input: command + current aggregate state
    • Output: list of events or error
    • Pure function; no side effects
  4. Implement projector function:
    • Input: event + current read model
    • Output: updated read model
    • Also pure
  5. Orchestration engine:
    • Receives command from client
    • Validates via commandInvariants
    • Calls decider to produce events
    • Appends events to event log (durable storage)
    • Calls projector to update in-memory read model
    • Returns new projected state to client
  6. On startup, replay event log through projector to rebuild read model

Example

const decider = (command, state) => {
  if (command.type === 'thread.create') {
    if (state.threads.has(command.threadId)) return Err('thread exists')
    return Ok([{ type: 'thread.created', threadId: command.threadId, title: command.title }])
  }
  return Err('unknown command')
}

const projector = (event, readModel) => {
  if (event.type === 'thread.created') {
    readModel.threads.set(event.threadId, { id: event.threadId, title: event.title })
  }
  return readModel
}

const engine = new OrchestrationEngine()
const [events, newState] = yield* engine.dispatch(command)

Counter / Caveats

  • Event log grows unbounded; implement snapshot/compaction strategy
  • Projector must be deterministic; any randomness breaks replay
  • Decider can be slow for large aggregates; consider per-thread vs global state
  • Concurrent commands to same aggregate need optimistic locking or version checks

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.