AGENT SKILL

llm-providers

Configure transcript-based and video-understanding LLM clipping routes, provider prefixes, API keys, prompt format, and timestamp normalization for FunClip.

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

01 前言 Front Matter

name
llm-providers
license
MIT
metadata
{ "disco-role": "operating" }
description
Configure transcript-based and video-understanding LLM clipping routes, provider prefixes, API keys, prompt format, and timestamp normalization for FunClip.
disable-model-invocation
true

02 资源文件 4

03 SKILL.md 正文

LLM Providers

Use this sub-skill when the user needs to choose, configure, or debug a FunClip LLM backend.

Handles

  • funclip/launch.py llm_inference routing for transcript-based and video-understanding models.
  • funclip/llm/openai_api.py, qwen_api.py, litellm_api.py, g4f_openai_api.py, and twelvelabs_api.py.
  • The default prompt shape in funclip/llm/demo_prompt.py.
  • The timestamp format expected by funclip/utils/trans_utils.py::extract_timestamps.
  • AI Clip handoff: producing parseable LLM Clipper Result text for the clip-workflows sub-skill.

Route elsewhere

  • Transcript/video/audio state, clip execution, subtitle burn-in, offsets, and clip output files -> ../clip-workflows/SKILL.md.
  • Release archives, version packaging, and publishing -> ../release-packaging/SKILL.md.

Start here

  • Read references/providers.md for the provider table and key/env rules.
  • Read references/prompt-format.md for parseable N. [HH:MM:SS,mmm-HH:MM:SS,mmm] text output.
  • Read references/troubleshooting.md for auth, import, timeout, and timestamp failures.
  • Run scripts/provider_route_smoke.py --repo-root <repo-root> to check prefix stripping and Pegasus normalization without API calls.

Natural requests this covers

  • Set up GPT, Qwen, DeepSeek, AtlasCloud, MiniMax, LiteLLM, g4f, or TwelveLabs Pegasus for LLM clipping.
  • Diagnose missing API keys, wrong model prefixes, empty model names, or region/base-URL mismatches.
  • Make LLM output parseable by AI Clip.
  • Normalize Pegasus decimal-second timestamps into SRT form.
  • Decide when Pegasus needs an uploaded video instead of transcript-only input.

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