sdk
Skill2 starsChanged 5 months agoMITView on GitHub
Codex SDK 集成开发。用 TypeScript/Python SDK 做 agent 生命周期、流式输出、本地/云端执行等集成。
We split this skill into the rules it gives an agent
3 rules. 0 are common, 3 are distinctive, and 0 are very specific to this project.
0
Common
Most of the 10 edited versions we read keep these.
3
Distinctive
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 11
Use repository code and installed types first, then current primary documentation when needed: TypeScript SDK, Python SDK.
- Distinctive: line 21
Use supported disposal/context-manager patterns and await the terminal run result before claiming completion.
- Distinctive: line 25
Keep agent IDs and run IDs distinct.
“Common” is measured: kept by at least 6 in 10 of the 10 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Codex SDK
Build integrations with TypeScript @cursor/sdk or Python cursor-sdk / cursor_sdk when these are the requested or established packages. Preserve the user's selected library and runtime; identify the installed dependency and version before applying version-sensitive API details.
Use repository code and installed types first, then current primary documentation when needed: TypeScript SDK, Python SDK. This skill is guidance, not a substitute for the current package contract.
Choose the shape
Infer language from the request and target codebase. Ask only when the integration location or language remains materially ambiguous. For the selected SDK version, choose a one-shot prompt for one request, a persistent agent for streaming/follow-ups, or resume for an existing agent. Set the intended local/cloud runtime explicitly so the default cannot send work to the wrong environment.
Read API patterns and runtime details for implementation examples, lifecycle handling, failures, MCP, or later inspection. Load the relevant sections rather than treating every example as mandatory. Verify model IDs and availability from the actual model list; examples are not a recommendation to hardcode an old default.
Invariants to preserve
- Use supported disposal/context-manager patterns and await the terminal run result before claiming completion.
- Distinguish transport/startup errors, failed runs, cancellation, and unknown completion. Check run IDs/state before retrying when execution may have started; a retryable transport error alone is not proof an external side effect did not happen.
- Bound retries and honor backoff. Preserve user scope for cloud runs, PR creation, reviewer requests, and other external effects.
- Pass secrets through supported credential/environment mechanisms without printing them. Inspect actual ambient-settings behavior before relying on user machine defaults.
- Keep agent IDs and run IDs distinct. Verify supported operations on resumed/detached handles.
- Inline MCP configuration may need to be supplied again on resume, and per-send overrides can replace rather than merge servers; confirm the installed version's behavior.
Produce the working integration and run meaningful checks suited to its actual risk. Do not add an unrelated Canvas or an extra confirmation just because the integration contains structured output.
Other versions11 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 11
- kortix-ai/suna★ 20ksame name
- one-ie/one★ 132same name
- tinyhumansai/tiny.place★ 132same name
- ArangoGutierrez/claude-toolkit★ 4same name
- gabriellst/codm★ 4same name
- aiperceivable/apcore-skills★ 1same name
- Goodnessmbakara/codex-skills★ 1same name
- JameMy0001/antigravity-skills-hub★ 1same name
- leishifu666/skills★ 1same name
- filipefalcaofs/repo-padrao★ 0same name
1 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository57 other files it gives its agents.
- codex-skills-pack/AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- automateSkill
- autopilotSkill
- create-hookSkill
- create-subagentSkill
- goalSkill
- loopSkill
- onboardSkill
- subscribeSkill
- review-agentSkill
- review-bugbotSkill
- review-securitySkill
- reviewSkill
- split-to-prsSkill
- canvasSkill
- migrate-to-buildsSkill
- migrate-to-skillsSkill
- new-repoSkill
- originSkill
- rename-chatSkill
- shareSkill
- shellSkill
- statuslineSkill
- update-cli-configSkill
- update-cursor-settingsSkill
- computer-useSkill
- documentsSkill
- excel-live-controlSkill
- google-docsSkill
- google-drive-commentsSkill
- google-driveSkill
- google-sheetsSkill
- google-slidesSkill
- pdfSkill
- PresentationsSkill
- sites-buildingSkill
- sites-hostingSkill
- sites-preview-troubleshootingSkill
- SpreadsheetsSkill
- template-creatorSkill
- create-ruleSkill
- create-skillSkill
- imagegenSkill
- openai-docsSkill
- plugin-creatorSkill
- plugin-managementSkill
- skill-creatorSkill
- skill-installerSkill
- humanizer-zh-academicSkill
- natural-fictionSkill
- natural-writingSkill
- precise-proseSkill
- prose-auditSkill
- qu-ai-weiSkill
- research-paper-writingSkill
- voice-profileSkill
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
Reviews can't be read right now.
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.

