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/ideoshi/glyphloop/claude-mdgit clone --depth 1 https://github.com/ideoshi/glyphloopWhat 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 | $0.00722 | $0.00722 |
| Opus 5 | $0.00361 | $0.00361 |
| Sonnet 5 | $0.00144 | $0.00144 |
| Haiku 4.5 | $0.00072 | $0.00072 |
Grade A, and why
glyphloop 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 yesterday.
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 — 62 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Glyphloop
Local web app + headless CLI for loop-perfect generative ASCII animations (web embeds, MP4/GIF/PNG, terminal players). Imported video repeats one sampled pass and is seamless only when the source footage already loops.
Commands
npm run dev- editor UI on http://localhost:5199 (browser preview via.claude/launch.json)npm test- vitest unit tests (always run before committing)npm run build- typecheck + production buildnpm run render -- --preset <file> --out <dir>- headless render; see AGENTS.mdnpm run mcp- start the MCP server (stdio) so an agent can drive renders
Generating animations as an agent
Read AGENTS.md first - it has the preset JSON schema, per-source parameter tables, MCP tool reference, and embed integration snippets.
Headless sources (render via CLI/MCP): flowfield, waves, blobs, rain, shapes3d, expr, parametric3d. Image/video/text base layers and PNG/GIF/MP4 export are browser-editor features (canvas/codecs) and cannot render headless.
To expose the MCP server to an agent from any project, register the repo-local binary (replace the path with this repo's absolute path):
claude mcp add --scope user glyphloop -- node "/absolute/path/to/glyphloop/bin/glyphloop.js" mcp
Use the public npx glyphloop@beta render ... and npx glyphloop@beta mcp
commands from any project. For development against this checkout, use
npm run render -- ... and npm run mcp from inside the repo.
Architecture
Source (pure fn of t, seed, params - must be loop-periodic)
→ FieldFrame (src/core/field.ts, brightness grid)
→ AsciiMapper (src/core/mapper.ts, brightness → char + color)
→ CanvasRenderer (src/core/render.ts)
→ exporters (src/export/*)
src/core/state.ts-AppState, the single serializable state = preset format. Keep it import-clean: no UI or CJS-dep imports (the CLI loads it under Node ESM).src/sources/- one file per source +SOURCESregistry inindex.ts.src/cli/preset.ts- validates presets and rejects browser-only bases/layers (image/video/text) so the headless CLI only accepts renderable input.src/cli/- preset merge/validation + render command (run via tsx).src/mcp/server.ts- MCP server (list_sources, preview_frame, render_animation).- Loop-perfection rule: time may only enter a source via
loopCoords()noise sampling or integer-cycle2π·k·t/Tphases. Tests enforce t=0 ≡ t=duration.
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.
- yesterday First seen · 62 lines · 722 tokens per session scan A b37bdc9540d6
glyphloop CLAUDE.md is an instructions file published in the GitHub repository ideoshi/glyphloop (0 stars, last pushed 1mo ago), licensed MIT. It adds 722 tokens to every session, about $0.0036 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 2026-08-31.
Other instructions, from other repositories
Paperlab AGENTS.md
Instructions for NourMtir0722/Paperlab, covering paperlab — for coding agents, integrating paperlab into a project, stage mode — paper as architecture, content types and lighting is data, not an enum.
markdy-com copilot-instructions.md
Instructions for HoangYell/markdy-com, a project described as: 🎬 Open-source animated architecture diagrams as code. Diagram-native DSL, 17 layout engines, Web Animations API, and MCP server for AI agents.
ShipSwift AGENTS.md
Instructions for signerlabs/ShipSwift, covering agents.md, project overview, directory structure, naming conventions and dependency rules.
ssgoi AGENTS.md
Instructions for meursyphus/ssgoi, covering ssgoi and releasing.
Unity-AI-Animation CLAUDE.md
Instructions for IvanMurzak/Unity-AI-Animation, covering claude.md, what this is, build / run, critical invariants and find detail in.
deadline-cloud AGENTS.md
Instructions for aws-deadline/deadline-cloud, covering agents.md - aws deadline cloud client, project overview, repository structure, key commands and development.