Agent skill · data analytics · jeremylongshore

klingai-job-monitoring

Track and monitor Kling AI video generation task status. Use when building

Why this skill is useful

Provides specific polling strategies and error handling for the Kling AI video generation API that the AI wouldn't reliably generate on its own.

What it needs

Requires npm installed locally. Requires klingai account access. About 3k tokens when loaded. Last updated 2026-08-07. 2,604 stars on the source repository.

What this skill does

Kling AI Job Monitoring Overview Every Kling AI generation returns a taskid. This skill covers polling strategies, batch tracking, timeout handling, and callback-based monitoring for the /v1/videos/text2video, /v1/videos/image2video, and /v1/videos/video-extend endpoints. Task Lifecycle Status Meaning Typical Duration -------- --------- ----------------- submitted Queued for processing 0-30s processing Generation in progress 30-120s (standard), 60-300s (professional) succeed Complete, video URL available Terminal failed Generation failed Terminal Polling a Single Task Batch Job Tracker Stuck Task Detection Batch Monitor Loop Prerequisites An approved job queue, synthetic or rights-cleared briefs, an authorized workspace and credit cap, draft-only destination, policy review, and cancellation/removal owner. Instructions 1. Monitor only approved sandbox or production-canary tasks; store task references and aggregate state counts, not prompts, asset URLs, or identities. 2. Verify task ownership, policy/rights status, credit consumption, retention, and draft-only routing before any downstream publication step. 3. Pause and cancel queued tasks on stuck jobs, unexpected cost, policy, rights, scope, or retention drift; remove associated temporary drafts. 4. Keep a redacted monitoring receipt for the approved window and ensure a named owner can restore the prior queue configuration. Output Produce a monitoring receipt with environment, aggregate task states, queue limits, credit use, policy/rights/draft-only checks, cancellation outcome, owner, retention/removal proof, and rollback reference. Exclude prompts, assets, and credentials. Error Handling Condition Response --- --- Stuck or duplicate task Pause the queue, cancel or deduplicate the task, and investigate using redacted metadata only. Policy, rights, budget, or retention drift Cancel affected drafts, remove temporary assets, and require owner review before resuming. …

How to use it

Reference it in AdaL, Claude Code, Cursor or any coding agent — nothing to install:

@skills jeremylongshore/klingai-job-monitoring--b57fcd

View the source on GitHub

Browse the @skills marketplace