AGENT SKILL

camera-recording-pipeline

Guides Viseron camera ingestion, stream command construction, recorder/NVR behavior, storage tiers, recordings, snapshots, timelapse, live relay, and camera/NVR troubleshooting.

License 许可
MIT
最近更新
2026-09-03
Stars
★ 351
Forks
27

01 前言 Front Matter

name
camera-recording-pipeline
license
MIT
metadata
{ "disco-role": "operating" }
description
Guides Viseron camera ingestion, stream command construction, recorder/NVR behavior, storage tiers, recordings, snapshots, timelapse, live relay, and camera/NVR troubleshooting.
disable-model-invocation
true

02 资源文件 5

03 SKILL.md 正文

Camera Recording Pipeline

Use this sub-skill when the task is to configure or diagnose Viseron camera streams, FFmpeg/GStreamer command behavior, NVR coordination, recorder settings, retention tiers, snapshots, timelapse output, or live camera viewing.

Route by task

  • Camera source, URL, FFmpeg/GStreamer, substream, FFprobe, hardware acceleration, raw command, or recoverable decode errors: read camera streams. Use scripts/build_ffmpeg_url.py to safely construct and redact camera URLs without contacting a camera.
  • NVR coordination, event/continuous/manual recordings, recorder lookback and segments, storage tiers, snapshots, timelapse, event clips, or cleanup safety: read storage and recordings.
  • Live View, MJPEG endpoints, static MJPEG streams, go2rtc/WebRTC/MSE, or restreaming through go2rtc: read live view and go2rtc.
  • Camera/NVR/storage/live-view failure diagnosis: read troubleshooting before changing config.

Boundaries

This sub-skill owns camera ingestion, stream command construction, recorder/NVR operation, retention and cleanup planning, snapshots/timelapse storage, live camera relay, and camera/NVR failure modes.

Route detector labels, masks, zones, detector model/device choices, trigger_event_recording semantics beyond its recorder effect, and post-processor logic to detection-and-ai-components. Route Docker installation, global config.yaml/secrets.yaml, webserver/authentication/logging, and safe-mode startup to configuration-and-deployment. Route MQTT, webhooks, notifications, Telegram commands, Home Assistant discovery, and PTZ control surfaces to automation-and-integrations. Route component/domain implementation, tests, reload behavior, and source-code changes to developer-extension.

Safe workflow defaults

  1. Build or redact URLs locally; never print real camera passwords unless the user explicitly asks for raw output.
  2. Treat live cameras, hardware acceleration, GStreamer system libraries, go2rtc ports, database services, containers, and network shares as target-host requirements unless separately verified on that host.
  3. Prefer generated FFmpeg/GStreamer commands before using raw_command/raw_pipeline; raw overrides skip many safety defaults.
  4. Keep recorder and snapshot retention separate: recorder tiers do not protect snapshots, and snapshot tiers do not protect event video.
  5. Do not run destructive database reset helpers as a cleanup shortcut; use retention rules and normal Viseron cleanup behavior first.

04 同仓库更多 Skills

VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-triton-opt

Make a single PassNet Triton kernel fast AND numerically faithful: performance model (when a replacement can win at all), block/grid/warp tuning, autotune policy, launch overhead, and per-op numeric recipes to pass the dtype baseline tolerances. Use when a pass matches and is correct but speedup below expected, or when correctness fails by small numeric margins.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-solve

END-TO-END playbook for solving one PassNet sample: analyze the computation graphs, decide the optimization strategy (what to fuse, what to replace, what to leave alone), drive the iteration loop, and maximize the sample score. This is the ENTRY skill — invoke it first for any PassNet optimization task; it tells you when to use passnet-pattern-fusion, passnet-triton-opt and passnet-feedback.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-skill

PassNet GPU kernel optimization via compiler passes. Design and implement Triton-based optimization passes, create pass files under ./pass_dir/, self-evaluate with pass_evaluator, and iterate to maximize GPU speedup.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-pattern-fusion

Author PassNet pass files: write patterns that actually MATCH the FX graph, pick fusion regions (which ops to absorb into one Triton kernel), and structure multi-pass submissions with the mandatory shared-dispatch architecture. Use when creating or fixing pattern()/replacement_args()/replacement_func() files, when a pass "failed to match", or when deciding how to fuse multiple kernels/ops.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-orchestrate

Round/batch planning and decision support for PassNet graph optimization. Use for multi-sample triage, worker allocation, eval-budget policy, and post-eval keep/revert decisions. For solving one concrete sample end to end, use passnet-solve as the entry skill; consult this skill only when you need broader planning or a specific decision gate.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-feedback

Fast, deterministic feedback for PassNet work: pre-flight pattern/match verification WITHOUT burning a GPU evaluation, per-node bottleneck analysis of a sample's graphs, and eval-log parsing into per-variant status + estimated score + failure classification. Use BEFORE every GPU evaluation (check_pattern), at the START of a sample (analyze_graph), and AFTER every evaluation (parse_eval_log).

★ 351