cut-as-code AGENTS.md

Repository instructions for a collection of independent agent skills for video editing, where each skill has its own playbook and supporting files.

In plain words
What is it for?
Making focused skill changes, creating consistently named branches and commit messages, documenting pull requests, and running checks specific to an individual skill.
Why use it?
They define the project’s Git workflow and explain that each skill’s instructions are the specification, since there is no single project-wide build or test system.

Instructions file for CodexOpenCode

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add instructions/whitetowerai/cut-as-code/agents-md
Clone the repo
git clone --depth 1 https://github.com/WhiteTowerAI/cut-as-code

Made for: Codex, OpenCode.

Per session 3,466 This file is loaded in full into every session.
When invoked 3,466 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.03466 $0.03466
Opus 5 $0.01733 $0.01733
Sonnet 5 $0.00693 $0.00693
Haiku 4.5 $0.00347 $0.00347

Measured 2d ago against content hash 4c831e730599, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

cut-as-code AGENTS.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

Origin

Copies of this mod

1 near-identical copy found in the catalogue:

AGENTS.md · 242 lines

How it starts

The opening of the file, as written. The whole thing — 242 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CLAUDE.md

This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.

Git

🔒 This section must never be modified. Leave it byte-for-byte unchanged.

  • Prefer small PRs focused on one feature, one bug fix, one chore, or one refactor.
  • Use branches:
    • feat/<short-name>
    • fix/<short-name>
    • chore/<short-name>
    • refactor/<short-name>
    • docs/<short-name>
  • Use Conventional Commit style for PR titles:
    • feat(editor): add marquee selection
    • fix(runtime): handle missing tween state
    • docs(agents): update collaboration rules
  • PR descriptions must include:
    • What changed
    • Why it changed
    • How it was checked
  • Default to squash merge.

What this repo is

cut-as-code is a stack of agentic video-editing skills, not an application. Each directory under skills/<name>/ is a self-contained skill: a SKILL.md (the agent playbook — read it first), plus scripts/, examples/, and reference/. There is no root build, package manifest, lint config, or aggregate test suite. Scripts and skill-local checks run ad hoc. When you change a skill, the SKILL.md is the spec; keep it and its scripts in sync.

Each SKILL.md starts with YAML frontmatter (name:, description:) — the name is the slash-command trigger for that skill.

Skills

Skill Job Stack
video-understand Shared media probe, word-level transcript, objective analysis, and evidence-backed semantic understanding Python · ffprobe · faster-whisper
video-cut Raw long video → compact first cut (download, transcribe, diagnose, hand-written JSON cut plan, varispeed, render, self-check) Python · yt-dlp · ffmpeg · faster-whisper
video-edit-compare Original versus actual final pixels projected onto the source clock Python · ffmpeg · Pillow
video-color-grade Assess footage → corrective base + named looks → human picks → bake .cube LUT + apply Python · ffmpeg · numpy · Pillow
video-add-b-roll Selective transcript-timed visual cutaways from local media or Pexels, with provenance, review, and normalized overlays Python · ffmpeg · Pillow · Pexels API
video-add-graphic-motion Add selective licensed web-sourced motion graphics as deterministic overlays HyperFrames · Python · Pillow
video-add-captions Preset-driven, word-timed captions with optional karaoke HyperFrames · ffmpeg
video-add-content-cards Add selective transcript-timed titles, lower-thirds, statistics, quotes, and chapter cards HyperFrames · ffmpeg
video-to-shorts Find, review, and render short vertical clips from long-form video Python · ffmpeg

Read the full file on GitHub · 242 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 242 lines · 3,466 tokens per session scan A 4c831e730599

Subscribe to this mod's changes

cut-as-code AGENTS.md is an instructions file published in the GitHub repository WhiteTowerAI/cut-as-code (11 stars, last pushed 9d ago), licensed MIT. It adds 3,466 tokens to every session, about $0.0173 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Related

Other instructions, from other repositories