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/elliotttate/splicekit/agents-mdgit clone --depth 1 https://github.com/elliotttate/SpliceKitWhat 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 | $0.09879 | $0.09879 |
| Opus 5 | $0.04940 | $0.04940 |
| Sonnet 5 | $0.01976 | $0.01976 |
| Haiku 4.5 | $0.00988 | $0.00988 |
Grade A, and why
SpliceKit 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.
Copies of this mod
1 near-identical copy found in the catalogue:
- SpliceKit CLAUDE.md — 86% identical, 215 lines differ
How it starts
The opening of the file, as written. The whole thing — 906 lines — stays where its author put it; the contents beside it link to each section on GitHub.
SpliceKit - Programmatic Final Cut Pro Control
SpliceKit is an ObjC dylib injected into FCP's process. It exposes all 78,000+ ObjC classes via a JSON-RPC server on TCP 127.0.0.1:9876. Everything is fully programmatic -- no AppleScript, no UI automation, no menu clicks.
NEVER Use Keyboard Simulation or AppleScript
All FCP actions MUST go through direct ObjC calls. Never use any of these:
SpliceKit_simulateKeyPress()or synthetic NSEvent key eventsCGEventkeyboard/mouse postingosascript/NSAppleScript/ AppleScripttell application- Accessibility APIs (
AXUIElement) for clicking or typing - Frame-stepping loops (
nextFramex N) whenseekToTimeexists
Instead, always use one of these patterns (in priority order):
- Direct method call on the timeline/sequence/clip object via
objc_msgSend - Responder chain via
[NSApp sendAction:NSSelectorFromString(@"selector:") to:nil from:nil] - Batch bridge endpoint (e.g.
timeline.addMarkers) for operations on multiple items - Menu execute via
menu.executefor actions only accessible through menus
If an action doesn't have a direct ObjC selector yet, find one using explore_class /
search_methods on the relevant class (FFAnchoredTimelineModule, FFAnchoredSequence, etc.)
rather than simulating the keyboard shortcut.
Quick Start
1. bridge_status() -- verify connection
2. get_timeline_clips() -- see timeline contents
3. timeline_action("blade") -- edit
4. verify_action("after blade") -- confirm
CRITICAL: Must Know Before Editing
Opening a Project
If get_timeline_clips() returns an error about "no sequence", load a project:
# Navigate: library -> sequences -> find one with content -> load it
libs = call_method_with_args("FFLibraryDocument", "copyActiveLibraries", "[]", true, true)
libs_handle = json.loads(libs)["handle"]
lib = call_method_with_args(libs_handle, "objectAtIndex:", '[{"type":"int","value":0}]', false, true)
lib_handle = json.loads(lib)["handle"]
seqs = call_method_with_args(lib_handle, "_deepLoadedSequences", "[]", false, true)
seqs_handle = json.loads(seqs)["handle"]
allSeqs = call_method_with_args(seqs_handle, "allObjects", "[]", false, true)
# For each sequence result, extract its "handle" before calling hasContainedItems
# Example: seq_handle = json.loads(seq)["handle"]
# Check each: call_method_with_args(seq_handle, "hasContainedItems", "[]", false)
# Load: get NSApp -> delegate -> activeEditorContainer -> loadEditorForSequence:
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 · 906 lines · 9,879 tokens per session scan A 45cdf03bb54e
SpliceKit AGENTS.md is an instructions file published in the GitHub repository elliotttate/SpliceKit (131 stars, last pushed 3d ago), licensed MIT. It adds 9,879 tokens to every session, about $0.0494 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
fcp-mcp-server CLAUDE.md
Claude Code instructions for DareDev256/fcp-mcp-server, covering claude.md — fcp-mcp-server, what this is, architecture, key patterns and running.
diffusers AGENTS.md
AGENTS.md instructions for huggingface/diffusers, a project described as: 🤗 Diffusers: State-of-the-art diffusion models for image, video, and audio generation in PyTorch.
comfy-prompt-studio AGENTS.md
Instructions for yxhpy/comfy-prompt-studio, covering agents.md - ai 代理配置文档, ai 提供商, 1. ollama (默认), 2. gemini and 提示词生成服务.
davinci-resolve-mcp AGENTS.md
Instructions for apvlv/davinci-resolve-mcp, covering project knowledge base, structure, where to look, code map and conventions.
lianhuanhua-skills AGENTS.md
Instructions for littlewindy123/lianhuanhua-skills, covering agents.md, project goal, architecture rules, commands and before committing.
docvideoer CLAUDE.md
Instructions for coding-ax/docvideoer, covering claude.md — docvideoer, 触发方式, 外部 skill 依赖, 工作目录 and 核心流水线(6 步).