Agent skill · software engineering · jeremylongshore
klingai-ci-integration
Integrate Kling AI video generation into CI/CD pipelines. Use when automating
Why this skill is useful
Adds a complete CI/CD integration script for automating video generation with Kling AI that isn't available in public documentation.
What it needs
Requires pip, python installed locally. Requires klingai account access. About 4k tokens when loaded. Last updated 2026-08-07. 2,604 stars on the source repository.
What this skill does
Kling AI CI Integration Overview Automate video generation in CI/CD pipelines. Common use cases: generate product demos on release, create marketing videos from prompts in a YAML file, regression-test video quality across model versions. GitHub Actions Workflow CI Generation Script Batch from YAML Config GitLab CI Secret Management Platform Store AK/SK in ---------- --------------- GitHub Actions Repository Secrets GitLab CI CI/CD Variables (masked) AWS CodeBuild Parameter Store / Secrets Manager GCP Cloud Build Secret Manager Never put API keys in the workflow YAML or commit them to the repo. Prerequisites A CI environment with Python 3.11+, pinned dependencies, a secret-manager-backed Kling credential, and an explicit per-run credit and concurrency budget. A repository-controlled model, duration, destination, and content-policy allowlist. CI fixtures must be synthetic or rights-cleared; never use customer media or real-person likenesses in unattended jobs. A private artifact bucket, short retention period, and an approval gate. Automated jobs produce draft, watermarked media only until a named owner approves promotion. Instructions 1. Validate the workflow and prompt manifest before any API call: require a synthetic/rights-cleared fixture identifier, approved model and duration, permitted destination, and a nonzero but bounded budget. 2. Load credentials only from masked CI secrets, run a policy and consent check, and use a stable manifest hash to prevent duplicate submissions on retries. 3. Submit a single sandbox canary first. Assert that output remains private and watermarked, that no source or prompt is echoed into logs, and that the run has not exceeded its credit or concurrency budget. 4. Require an owner approval artifact before promotion. Publish by immutable output digest to the allowlisted bucket; do not publish directly from a provider URL. 5. …
How to use it
Reference it in AdaL, Claude Code, Cursor or any coding agent — nothing to install:
@skills jeremylongshore/klingai-ci-integration--eab865