Temporal orchestration

Skill yigityildiz0/universal-ai-skill-library/skills/common/temporal-orchestration

531 searchable AI Agent Skills for Claude Code, OpenAI Codex, and OpenCode — EN/TR catalog, platform and risk notes, direct ZIPs, and curated bundles.

Install

npx -y skills add yigityildiz0/universal-ai-skill-library --skill temporal-orchestration

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

3 things to look at

19 days oldThe repository was created 19 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
1 stars1 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

Design, implement, and verify durable Temporal workflows and activities for long-running or failure-prone processes. Use for retries, timers, signals.

The file declares its own license as MIT. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.

SKILL.md

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

Temporal Orchestration

Boundary

Workflow code must be deterministic and replay-safe. Put network, filesystem, database, randomness, clock access, and AI/provider calls in activities. Never place credentials or large payloads in workflow history.

Design

Define business ID, idempotency key, inputs/outputs, states, timeouts, cancellation, retention, and completion criteria.
Split activities around independently retryable external effects, not every function call.
Configure schedule-to-close/start-to-close/heartbeat timeouts and retry policy from failure semantics. Mark validation, authorization, and permanent business failures non-retryable.
Make activities idempotent because at-least-once execution is possible. Record external operation IDs before retryable effects.
Use signals for commands, queries for read-only state, updates when a validated synchronous mutation is required, and child workflows for independently managed lifecycles.
Model compensation explicitly; compensation is not a database rollback and must itself be retryable/idempotent.
Bound history growth with continue-as-new and pass compact state, not an ever-growing event log.

AI activities

Keep prompt/model/provider selection and billing/privacy controls inside activity configuration. Use the project's current configured provider/model and verify support at runtime; do not hard-code a default or change it automatically. Persist only the minimum reproducible request/result metadata permitted by privacy policy. Bound tokens, tool rounds, retries, concurrency, and cost. Do not retry non-idempotent tool effects blindly.

Evolution and deployment

Use Temporal's supported workflow versioning/deployment mechanism for the installed SDK/server version. Existing executions must replay against compatible code. Test replay of real histories before removing old paths. Deploy workers with graceful shutdown, task-queue compatibility, observability, and rollback.

Testing

Run deterministic unit tests, time-skipping tests, activity mocks, retry/timeout/cancellation tests, signal/update races, duplicate activity execution, compensation failure, continue-as-new, worker restart, and history replay. Add a bounded integration test against an authorized local/test server.

Completion report

Document workflow/activity boundaries, IDs, timeouts/retries, idempotency, history strategy, versioning, worker topology, tests, observability, security/data boundary, and rollback. Do not deploy or alter a production namespace without explicit authorization.