Borrowing it
Nothing to install: this file belongs to ChaosRealmsAI/NextFrame. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/ChaosRealmsAI/NextFrame/main/CLAUDE.mdgit clone --depth 1 https://github.com/ChaosRealmsAI/NextFrameWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/chaosrealmsai/nextframe/claude-md)<a href="https://agentmods.dev/instructions/chaosrealmsai/nextframe/claude-md"><img src="https://agentmods.dev/badge/instructions/chaosrealmsai/nextframe/claude-md.svg" alt="Measured on agentmods" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.01483 | $0.01483 |
| Opus 5 | $0.00741 | $0.00741 |
| Sonnet 5 | $0.00297 | $0.00297 |
| Haiku 4.5 | $0.00148 | $0.00148 |
Grade A, and why
NextFrame CLAUDE.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 8d 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.
How it starts
The opening of the file, as written. The whole thing — 68 lines — stays where its author put it; the contents beside it link to each section on GitHub.
NextFrame · AI Entry
NextFrame is an AI-native video editor/runtime: structured JSON compositions render in the desktop editor and export to MP4.
Run
# First-time setup · build the frontend bundle (gitignored · webview loads it at runtime)
cd frontend/nf-components && npm install && npm run build && cd -
cargo build -p nf-shell -p nf-cli
target/debug/nf-shell
target/debug/nf open --project v2-showcase --composition showreel-clip-first
Example projects live under examples/; runtime projects are copied to ~/.nextframe/.
Black preview gotcha · if the desktop window shows the idle "NextFrame · AI-FIRST CLI-DRIVEN" hero
instead of your composition, frontend/nf-components/dist/index.js is missing — re-run the npm build above.
The webview silently falls back to idle when the bundle isn't loaded; nf devtools --eval='document.querySelectorAll("[data-nf-component]").length' returning 0 confirms it.
AI 验证接口
target/debug/nf open --project=<slug> --composition=<slug> — open a composition in the desktop editor.
target/debug/nf screenshot --project=<slug> --episode=<slug> --region=editor --out=<png> — probe a desktop editor region and write a size stub PNG; use capture for visual evidence.
target/debug/nf capture --project=<slug> --episode=<slug> --out=<png> — capture the native macOS window PNG for real visual verification.
target/debug/nf click --project=<slug> --episode=<slug> --selector=<css> — simulate a real desktop click, including shadow DOM selectors.
target/debug/nf devtools --project=<slug> --episode=<slug> --query=<css> --get=<prop> — inspect live DOM, including shadow DOM selectors.
target/debug/nf devtools --project=<slug> --episode=<slug> --query=<css> --fill=<value> [--get=<prop>] — fill a live input through the same input/change path as human editing.
target/debug/nf devtools --project=<slug> --episode=<slug> --eval=<js> — evaluate JavaScript in the live desktop webview and return {eval,value,error} JSON.
target/debug/nf-shell 2>&1 | grep NFCONSOLE — watch forwarded webview console.log/warn/error/info lines from the shell process.
target/debug/nf cue --timeline=<vox.json> [--max-chars=18] [--min-pause-ms=250] [--out=<json>] — use an LLM to cut a vox word-level timeline into validated cue list JSON.
target/debug/nf composition show --project=<slug> --composition=<slug> [--clip=<id>] [--track=<id>] [--item=<id>] [--field=<path>] — read raw composition JSON or one clip/track/item field.
target/debug/nf composition patch --project=<slug> --composition=<slug> [--clip=<id>] --track=<id> [--item=<id>] --field=<path> --value=<json-or-string> — patch one track or clip item field such as params.title, style.x, or time.start.
target/debug/nf composition validate --project=<slug> --composition=<slug> — validate component registry, files, mount/update exports, import-free ABI, used tracks, and observed params.
target/debug/nf composition compile --project=<slug> --composition=<slug> --out=<render_source.json> — compile AI-authored composition JSON into stable nf.render_source.v1 recorder input.
target/debug/nf verify --project=<slug> --composition=<slug> [--out=<json>] [--screenshot-dir=<dir>] — verify AI-authored composition JSON: component ABI, compiled source, overlap intent, anchor guide, ASCII timeline, layout/text checks, and typical screenshot commands.
target/debug/nf-recorder validate-source --source=<render_source.json> — validate recorder input contract without reading project storage or desktop state.
target/debug/nf-recorder export --source=<render_source.json> --profile=draft|standard|final|final-fast --output=<mp4> [--diagnostics=<json>] — export directly from render source through the recorder module.
target/debug/nf-recorder snapshot-source --source=<render_source.json> --t-ms=<ms> --output=<png> — sample one render-source frame through the same recorder HTML/runtime path.
target/debug/nf export --project=<slug> --composition=<slug> --profile=draft|standard|final|final-fast --out=<mp4> — export a composition to MP4 with a named quality/speed profile.
target/debug/nf export --project=<slug> --composition=<slug> --profile=draft --diagnostics --out=<mp4> — export and write a sibling diagnostics JSON with frame timings, slow spans, and top slow frames.
target/debug/nf export --project=<slug> --composition=<slug> --fps=30|60 --resolution=720p|1080p|4k --parallel=<1-8> --events --out=<mp4> — override export settings and stream recorder progress JSONL before the final summary JSON.
target/debug/nf verify-export --source=<render_source.json> --video=<mp4> --out=<report.json> — sample exported MP4 clip frames and fail on magenta-background or blank-frame regressions.
target/debug/nf export-status --job-id=<id> — read desktop export job status, including progress and diagnostics summary/path when available.
target/debug/nf export-cancel --job-id=<id> — cancel a running desktop export job and stop its recorder process group.
./scripts/check-structure.sh — verify repository skeleton: root allowlist, no nested spec git, no tracked generated artifacts.
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.
- 8d ago First seen · 68 lines · 1,483 tokens per session scan A 5ec237acbc84
NextFrame CLAUDE.md is an instructions file published in the GitHub repository ChaosRealmsAI/NextFrame (24 stars, last pushed 4mo ago), licensed MIT. It adds 1,483 tokens to every session, about $0.0074 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.
Other instructions, from other repositories
codex AGENTS.md
AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.
vscode buildNext.instructions.md
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
next.js AGENTS.md
AGENTS.md instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.
langchain AGENTS.md
AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, applying the notice (cutover) and fallback chain (never fail the build).
spec-kit AGENTS.md
AGENTS.md instructions for github/spec-kit, covering agents.md, about spec kit and specify, quickstart — add a new integration in 5 steps, integration architecture and integrationmanifest — file tracking.