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/natolambert/colloquium/claude-mdgit clone --depth 1 https://github.com/natolambert/colloquiumWrote 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/natolambert/colloquium/claude-md)<a href="https://agentmods.dev/instructions/natolambert/colloquium/claude-md"><img src="https://agentmods.dev/badge/instructions/natolambert/colloquium/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 | $0.00953 | $0.00953 |
| Opus 5 | $0.00477 | $0.00477 |
| Sonnet 5 | $0.00191 | $0.00191 |
| Haiku 4.5 | $0.00095 | $0.00095 |
Grade A, and why
colloquium 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 3d 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 — 89 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Colloquium — Agent Instructions
What This Is
Colloquium is an agent-native slide creation tool for research talks. Markdown-based, git-friendly, AI-drivable.
Commands
- Use
uvfor all Python operations uv pip install -e .to install in dev modeuv run colloquium build examples/hello/hello.mdto build slidesuv run colloquium serve examples/hello/hello.mdfor dev server (with live reload)- To preview built slides, open the
.htmlfile directly (open examples/hello/hello.html), NOT via the serve URL uv run colloquium capture examples/hello/hello.mdto capture slides as PNGsuv run pytestto run tests
Changelog
- Update
CHANGELOG.mdwith every PR — one line per PR, not per commit - Format:
- Short summary of all changes in the PR ([#N](https://github.com/natolambert/colloquium/pull/N)) - Add under
## [Unreleased]
Architecture Patterns
CSS
- Slides use
display:noneby default,display:flexon.active - CSS specificity:
.slide--type.activenot.slide--type(avoids display override bugs) - CSS custom properties (
--colloquium-*) for all theme colors/fonts - Humans can override with custom CSS or per-slide
styledirectives
Navigation (presentation.js)
- Keyboard: arrows, Home/End, Space, PgUp/PgDn, F (fullscreen), Esc
- Click zones: left 1/3 back, right 2/3 forward
- Touch swipe support
- Hash-based URL routing (#1, #2, ...)
- Slide counter + progress bar
Slide Directives
Per-slide configuration via HTML comments:
<!-- layout: two-column -->— slide layout<!-- class: highlight -->— CSS classes<!-- style: background: #1a1a2e -->— inline styles<!-- notes: Speaker notes here -->— speaker notes (hidden in presentation)
Output
- All output is a single self-contained HTML file (CSS/JS inlined)
- KaTeX (CDN) for LaTeX math
- highlight.js (CDN) for code syntax highlighting
PDF Export
- Primary:
@media printCSS makes all slides visible with page breaks, hides nav - Users can
Cmd+Pin any browser for a clean PDF - CLI convenience:
colloquium exporttries system Chrome--headless --print-to-pdf - Optional ghostscript compression if
gsis on PATH
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.
- 3d ago First seen · 89 lines · 953 tokens per session scan A 8d968c489252
colloquium CLAUDE.md is an instructions file published in the GitHub repository natolambert/colloquium (269 stars, last pushed 29d ago), licensed MIT. It adds 953 tokens to every session, about $0.0048 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 2026-09-01.
Other instructions, from other repositories
beamer-skill copilot-instructions.md
Instructions for Noi1r/beamer-skill, covering beamer slide workflow — github copilot instructions, core workflow, actions, default preamble and hard rules (non-negotiable).
ChineseResearchLaTeX AGENTS.md
AGENTS.md instructions for huangwb8/ChineseResearchLaTeX, covering 中国科研常用 latex 模板集 - 项目指令, 项目目标, 目录结构, 当前分层模型 and 核心工作流.
academic-ref-inserter copilot-instructions.md
Instructions for fuwenhao12/academic-ref-inserter, covering academic ref inserter — github copilot instructions, core commands and formats.
SolvingMicroDSOPs CLAUDE.md
Instructions for llorracc/SolvingMicroDSOPs, covering claude.md, project overview, build commands, latex → pdf and python dependencies.
math-beamer-slides-skill copilot-instructions.md
Instructions for YixuanYang00/math-beamer-slides-skill, covering math beamer slides instructions, actions, reference routing, core rules and workflow.
codex-stata-for-economists CLAUDE.md
Claude Code instructions for maxwell2732/codex-stata-for-economists, covering core principles, folder structure, commands, render the markdown/pdf report (quarto + stata engine) and quality score for a do-file (0–100).