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 skills add jimezsa/opencolab --skill block-diagramgit clone --depth 1 https://github.com/jimezsa/opencolabWrote 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/skills/jimezsa/opencolab/block-diagram)<a href="https://agentmods.dev/skills/jimezsa/opencolab/block-diagram"><img src="https://agentmods.dev/badge/skills/jimezsa/opencolab/block-diagram/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/jimezsa/opencolab/block-diagram"><img src="https://agentmods.dev/badge/skills/jimezsa/opencolab/block-diagram.svg" alt="Reviewed on agentmods" width="80" 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 | $0.00084 | $0.02055 |
| Opus 5 | $0.00042 | $0.01027 |
| Sonnet 5 | $0.00017 | $0.00411 |
| Haiku 4.5 | $0.00008 | $0.00205 |
Grade C, and why
block-diagram scanned grade C with 2 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 9d 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.
Downloads and executes remote codehighSupply chain
curl | sh runs whatever the server returns today, which is not necessarily what it returned when this was reviewed.
"script": "curl -fsSL https://d2lang.com/install.sh | sh -s --", Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
"script": "curl -fsSL https://d2lang.com/install.sh | sh -s --", How it starts
The opening of the file, as written. The whole thing — 197 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Block Diagram Skill
Use this skill when the user wants a block diagram that explains a system, model, pipeline, service, architecture, or workflow.
This is the deterministic path for architecture visuals in OpenColab. The source of truth is a D2 file, not an image-generation prompt. The default rendered style is D2 sketch mode, which gives a hand-drawn look while keeping the diagram deterministic and editable. Diagrams should stay compact, and arrows should remain unlabeled unless the label carries specific technical meaning.
Mission
Given a textual architecture description:
- Normalize it into components, groups, and directed relationships.
- Write a canonical
diagrams/<slug>.d2source file. - Render
diagrams/<slug>.svg. - Render
diagrams/<slug>.pngwhen the user wants a raster image or when the result should be sent back to Telegram. - Return a short explanation of the diagram and any important assumptions.
Prerequisites
d2is installed and available inPATH.bashis available inPATH.- The working directory is writable.
Required Inputs
- A system or architecture description.
- Optional audience: beginner, engineer, executive, paper figure, infra team, and so on.
- Optional must-include components, flows, or labels.
- Optional must-include formulas or equations.
- Optional preferred output name.
- Optional render style:
sketchorclean.
If the user does not provide an audience, assume engineer.
If the user does not provide an output name, derive a short slug from the system name.
If the user does not provide a render style, use sketch.
Hard Requirements
- Use this skill for block diagrams instead of defaulting to a free-form image model.
- Treat the
.d2file as the canonical artifact and the rendered image as a derived artifact. - Do not invent components, edges, protocols, or subsystems that were not implied by the request.
- Support optional formula blocks with
equation: |latexonly when a mathematically important transformation, loss, or objective would be materially harder to understand without an explicit equation. - Do not generate a formula just because the request mentions math, a model, or an algorithm; include one only when the diagram would otherwise lose important meaning.
- If one ambiguity blocks a faithful diagram, ask one targeted question. Otherwise proceed autonomously.
- Keep the first diagram readable and compact. If the architecture is too dense, split it into an overview diagram plus one focused detail diagram.
- Use
references/style-guide.mdfor layout, color semantics, naming, and grouping rules. - Use
references/patterns.mdwhen selecting the diagram structure. - Use
ml-llm-architecture-template.mdas an optional template for detailed ML, LLM, neural-network, transformer, training, or quantization architecture diagrams. - Use
references/validation.mdbefore returning the final result. - Use
scripts/render_d2_diagram.shto format, validate, and render the final diagram. - Default to sketch-style rendering. Only switch to clean rendering when the user explicitly asks for a polished, paper-ready, or non-sketch output.
- Default to unlabeled arrows. Add edge labels only when they convey specific information such as a protocol, artifact, or payload that the arrow alone would not communicate.
- Never use generic edge labels such as
input,output,data,result,something input, orsomething output. - OpenColab normally provides
OPENCOLAB_PROGRESS_FILEduring provider runs. When it is set and the task is long enough to justify updates, emit bounded JSON progress events for normalization, draft completion, render start, warnings, blockers, or final artifact creation when they help the user understand real progress. - If the rendered artifact should be sent to Telegram, emit a raw
@telegram-file {"kind":"document","file":"diagrams/<slug>.svg","caption":"optional"}or@telegram-file {"kind":"photo","file":"diagrams/<slug>.png","caption":"optional"}line on its own line with no backticks or code fences. Keep the JSON on one line, use a validkind(photofor images, neverimage/png/jpg), and on Windows write the path with forward slashes. - If PNG rendering is unavailable in the current environment, fall back to the SVG artifact and send it as a Telegram document instead of pretending the PNG exists.
What ships with it
6 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 9d ago First seen · 197 lines · 84 tokens per session scan C 2fd482bf4bed
block-diagram is a skill published in the GitHub repository jimezsa/opencolab (11 stars, last pushed 26d ago), licensed MIT. It adds 84 tokens to every session and 2,055 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it C with 2 findings (downloads and executes remote code, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
skin-creator
Create and apply a two-asset LobsterAI visual skin from the user's style description. Use only when the AI Skin Designer kit supplies the structured skinpack workflow marker; do not use for ordinary theme or image requests.
canvas-design
Create beautiful visual art in .png and .pdf documents using design philosophy. You should use this skill when the user asks to create a poster, piece of art, design, or other static piece. Create original visual designs, never copying existing artists' work to avoid copyright violations.
taiyi-ui-design
A design-planning guide for describing how an application's user interface should look and behave. It produces a UI-DESIGN.md document covering layouts, components, interactions, accessibility, and error states.
solidworks-vibecad
A planning guide for turning natural-language mechanical design requests into structured, parameterized SolidWorks plans that can be reviewed before execution.
vs-search-tuning
Use when a user asks an agent to evaluate or tune text search similarity for an existing Viking AI Search application and dataset.
vs-search
Search runtime and scene management: verify queries, inspect scenes, debug app readiness, and diagnose recall or scene-config issues.