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.
npx agentmods add instructions/ronak-create/fablecut/claude-mdgit clone --depth 1 https://github.com/ronak-create/FableCutWhat 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 | $0.08937 | $0.08937 |
| Opus 5 | $0.04469 | $0.04469 |
| Sonnet 5 | $0.01787 | $0.01787 |
| Haiku 4.5 | $0.00894 | $0.00894 |
Grade A, and why
FableCut 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 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.
How it starts
The opening of the file, as written. The whole thing — 563 lines — stays where its author put it; the contents beside it link to each section on GitHub.
FableCut — browser video editor, drivable by Claude Code
A production-style non-linear video editor (Premiere-style) that runs in the
browser. An AI agent edits videos by editing project.json (or calling the
REST API / MCP tools) — the open browser UI live-reloads within ~150 ms via SSE.
No build step, no npm dependencies.
This file is the master manual. Any model pointed at this document (or at
the fablecut_docs MCP tool, which returns it) has everything needed to fully
drive the editor.
MCP connection (preferred — works from any session, any directory)
Register the MCP server (mcp-server.js) once at user scope as fablecut:
claude mcp add -s user fablecut -- node "<path-to>/fablecut/mcp-server.js".
Every Claude Code session then has these tools:
fablecut_status— auto-starts the editor server, returns URL + project summary. Call first.fablecut_docs— returns this document (section: "…"returns only matching##sections).fablecut_get_project/fablecut_set_project— read / replace the timeline JSON.fablecut_get_project {compact:true}returns a one-line-per-clip summary instead.fablecut_patch_project— apply targeted ops (add/update/remove clip/media, set project fields) without round-tripping the document. Prefer this for edits.fablecut_import_media— copy a local file into./media/and register it.fablecut_analyze_reference— turn a reference video into an edit blueprint (shots, beats, BPM, energy, drop) + extract its music. See "Remake a reference video".
Token-efficient editing (important for agents)
Editing via full get→modify→set costs thousands of tokens per change. Cheaper:
- Plan from
fablecut_get_project {compact:true}(≈10× smaller than the JSON) andfablecut_status— fetch the full JSON only to inspect exact keyframes. - Edit with
fablecut_patch_projectops — send only what changes, e.g.{ops:[{op:"updateClip", id:"c_v2", set:{props:{filterPreset:"noir"}}}]}. It re-reads the latest document internally, so it is merge-safe by design (no CONFLICT dance) and never destroys concurrent UI tweaks. - Docs: request
fablecut_docs {section:"schema"}(or "Recipes", "Remake", …) instead of the whole manual; skip it entirely if the schema is already in context. - Media questions (duration, fps, size): read them from the registered media entries — don't shell out to ffprobe; the browser probes and writes them back.
- Batch related changes into ONE patch call (ops apply in order, one revision bump).
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.
- 2d ago First seen · 563 lines · 8,937 tokens per session scan A 6ad47c1febf4
FableCut CLAUDE.md is an instructions file published in the GitHub repository ronak-create/FableCut (643 stars, last pushed 4d ago), licensed MIT. It adds 8,937 tokens to every session, about $0.0447 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
openscreen AGENTS.md
Instructions for getopenscreen/openscreen, covering agents.md, setup commands, development principles, project layout and code style.
VChart AGENTS.md
Instructions for VisActor/VChart, covering vchart 工程基准, 编码习惯, 文档规范 and 优先级.
open-edit AGENTS.md
Instructions for veedstudio/open-edit, covering agents.md, to run a video, how the pieces relate, hard rules (do not drift — these protect output quality) and conventions.
erm AGENTS.md
Instructions for dougcalobrisi/erm, covering agents.md — erm, install, fallback (no uv on path), use — the iterate loop and tune — five knob clusters.
movi-player AGENTS.md
Instructions for MrUjjwalG/movi-player, covering agents.md — movi-player for ai coding assistants, 1. what this is, 2. package entry points, 3. architecture map and 4. the web component .
leclap AGENTS.md
Instructions for heristop/leclap, covering agents.md, project overview, repository layout, setup and commands.