Klingai Sdk PatternsSAFE
Model-agnostic agent-skills platform with a harness-free canonical layer, verified adapters, and the ccpi package manager. Explore at tonsofskills.com.
Overview
Model-agnostic agent-skills platform with a harness-free canonical layer, verified adapters, and the ccpi package manager. Explore at tonsofskills.com.
4f83675ca38aOBSERVED · 2026-10-09Host compatibility
What the documentation claims. We have not run a compatibility test.
| Host | Status | Notes |
|---|---|---|
| claude-code | mentioned |
What it tells the agent
The instruction file, verbatim from the audited commit — this is the text the model reads, and the surface the audit's instruction layer examines. Quoted here so you can judge it without cloning anything.
--- name: klingai-sdk-patterns description: 'Production SDK patterns for Kling AI: client wrapper, retry logic, async polling, and error handling. Use when building robust integrations. Trigger with phrases like ''klingai sdk'', ''kling ai client'', ''klingai patterns'', ''kling ai wrapper''. ' allowed-tools: Read, Write, Edit, Bash(npm:*), Grep version: 1.18.0 license: MIT author: Jeremy Longshore <[email protected]> tags: - saas - kling-ai - sdk - patterns compatibility: Designed for Claude Code --- # Kling AI SDK Patterns ## Overview Production-ready client patterns for the Kling AI API. Covers auto-refreshing JWT, typed request/response models, exponential backoff polling, async batch submission, and structured error handling. ## Python Client Wrapper ```python import jwt import time import os import requests from dataclasses import dataclass, field from typing import Optional @dataclass class KlingConfig: access_key: str = field(default_factory=lambda: os.environ["KLING_ACCESS_KEY"]) secret_key: str = field(default_factory=lambda: os.environ["KLING_SECRET_KEY"]) base_url: str = "https://api.klingai.com/v1" token_buffer_sec: int = 300 poll_interval_sec: int = 10 max_poll_attempts: int = 120 # 20 minutes max timeout_sec: int = 30 class KlingClient: """Production Kling AI client with auto-refreshing JWT.""" def __init__(self, config: Optional[KlingConfig] = None): self.config = config or KlingConfig() self._token = None self._token_expires = 0 @property def _headers(self) -> dict: now = int(time.time()) if now >= (self._token_expires - self.config.token_buffer_sec): payload = {"iss": self.config.access_key, "exp": now + 1800, "nbf": now - 5} self._token = jwt.encode(payload, self.config.secret_key, algorithm="HS256", headers={"alg": "HS256", "typ": "JWT"}) self._token_expires = now + 1800 return {"Authorization": f"Bearer {self._token}", "Content-Type": "application/json"} def _post(self, path: str, body: dict) -> dict: r = requests.post(f"{self.config.base_url}{path}", headers=self._headers, json=body, timeout=self.config.timeout_sec) r.raise_for_status() return r.json() def _get(self, path: str) -> dict: r = requests.get(f"{self.config.base_url}{path}", headers=self._headers, timeout=self.config.timeout_sec) r.raise_for_status() return r.json() def _poll_task(self, endpoint: str, task_id: str) -> dict: """Poll with exponential backoff until task completes.""" interval = self.config.poll_interval_sec for attempt in range(self.config.max_poll_attempts): time.sleep(interval) result = self._get(f"{endpoint}/{task_id}") status = result["data"]["task_status"] if status == "succeed": return result["data"]["task_result"] elif status == "failed": raise KlingGenerationError(result["data"].get("task_status_msg", "Unknown")) # Increase interval up to 30s max interval = min(interval * 1.2, 30) raise KlingTimeoutError(f"Task {task_id} did not complete in time") # --- Public API --- def text_to_video(self, prompt: str, **kwargs) -> dict: body = {"model_name": kwargs.get("model", "kling-v2-master"), "prompt": prompt, "duration": str(kwargs.get("duration", 5)), "aspect_ratio": kwargs.get("aspect_ratio", "16:9"), "mode": kwargs.get("mode", "standard")} if kwargs.get("negative_prompt"): body["negative_prompt"] = kwargs["negative_prompt"] if kwargs.get("cfg_scale") is not None: body["cfg_scale"] = kwargs["cfg_scale"] if kwargs.get("callback_url"): body["callback_url"] = kwargs["callback_url"] task = self._post("/videos/text2video", body) task_id = task["data"]["task_id"] if kwargs.get("wait", True): return self._poll_task("/videos/text2video", task_id) return {"task_id": task_id} def image_to_video(self, image_url: str, **kwargs) -> dict: body = {"model_name": kwargs.get("model", "kling-v2-1"), "image": image_url, "duration": str(kwargs.get("duration", 5)), "mode": kwargs.get("mode", "standard")} if kwargs.get("prompt"): body["prompt"] = kwargs["prompt"] task = self._post("/videos/image2video", body) task_id = task["data"]["task_id"] if kwargs.get("wait", True): return self._poll_task("/videos/image2video", task_id) return {"task_id": task_id} def extend_video(self, task_id: str, **kwargs) -> dict: body = {"task_id": task_id, "prompt": kwargs.get("prompt", ""), "duration": str(kwargs.get("duration", 5)), "mode": kwargs.get("mode", "standard")} result = self._post("/videos/video-extend", body) new_task_id = result["data"]["task_id"] if kwargs.get("wait", True): return self._poll_task("/videos/video-extend", new_task_id) return {"task_id": new_task_id} class KlingError(Exception): pass class KlingGenerationError(KlingError): pass class KlingTimeoutError(KlingError): pass ``` ## Usage ```python client = KlingClient() # Synchronous (waits for result) result = client.text_to_video( "A cat playing piano in a jazz club", model="kling-v2-6", mode="professional", duration=5, ) print(result["videos"][0]["url"]) # Fire-and-forget (returns task_id) task = client.text_to_video("Ocean waves at sunset", wait=False) print(f
Trust audit
SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | PASS |
| L1 | Static analysis of the code | PASS |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | PASS |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- none-observed
- Network
- none-observed
- Shell
- none-observed
- Dependencies
- pinned
- Secrets in source
- none-found
Findings (0)
No findings outside the package's declared scope.
Gates applied: no_behavioural_pass.
4f83675ca38afull audit observations/trust-audit/skill/jeremylongshore__klingai-sdk-patterns.json · Report an issue / request a re-scanAudit history
Every audit this skill has had.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-09 | 4f83675ca38a | SAFE | B | 89 | first audit |
Questions
What does the Klingai Sdk Patterns skill do?
Model-agnostic agent-skills platform with a harness-free canonical layer, verified adapters, and the ccpi package manager. Explore at tonsofskills.com.
Is Klingai Sdk Patterns safe to install?
The audit found nothing in the source that contradicts what it says it does, and graded it B (89/100). Grade A is held back for packages that have also passed a sandboxed behavioural run, which is why a clean skill reads B.
What can Klingai Sdk Patterns access on my machine?
The audit observed no filesystem, network or shell use at all in its source.
Which assistants does Klingai Sdk Patterns work with?
Its documentation mentions claude-code. That is what the text claims, not a compatibility test we ran.
How current is this page?
The grade is for one exact copy of the source (4f83675ca38a), read on 2026-10-09. The repository is watched, and a new audit runs when it changes — this is the first audit.