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 yarlson/yarstack --skill diagram-creategit clone --depth 1 https://github.com/yarlson/yarstackWrote 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/yarlson/yarstack/diagram-create)<a href="https://agentmods.dev/skills/yarlson/yarstack/diagram-create"><img src="https://agentmods.dev/badge/skills/yarlson/yarstack/diagram-create/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/yarlson/yarstack/diagram-create"><img src="https://agentmods.dev/badge/skills/yarlson/yarstack/diagram-create.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 | $0.00108 | $0.02367 |
| Opus 5.5 | $0.00043 | $0.00947 |
| Sonnet 5.5 | $0.00022 | $0.00473 |
| Haiku 4.5 | $0.00011 | $0.00237 |
Grade A, and why
diagram-create 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 today.
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 — 179 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Diagram Creation
Diagrams are code. A project keeps one diagrams.py that defines its
diagrams with the diagram package. One command checks every
diagram against the layout rules, writes the HTML pages with inline SVG,
checks in Chromium that all text fits, and writes one PNG for each SVG. The
package fixes every color, font, and size, so all diagrams share one style.
D is the diagram folder of the project, docs/diagrams/ by default.
1. Decide whether to draw
- Draw only when a reader learns more from the picture than from a paragraph or a table. A list of things is a table. A one-box diagram is a sentence.
- Charts of data (bars, lines, scatter) are out of scope.
- The best move is usually deletion. Each node is a distinct idea: merge two nodes that always travel together. Each arrow carries information: remove an arrow that the layout already shows. A diagram is done when nothing can be removed. Aim for a density of about 4 out of 10.
2. Set up the diagram folder
- If
Dalready has adiagrams.pyand adiagram/package, use them. Follow the README or AGENTS.md in that folder. Those project rules override this skill. - If the project has an older generator in this style (a script with an
Svgclass andzone,path,label, andnodehelpers), keep using it. It has none of the build checks. Port it to the package only when the user asks: each path string becomes a list of waypoints, and each label becomes thelabelof its edge. - Otherwise, copy
scripts/diagrams.pyand thescripts/diagram/package from this skill's directory intoD.diagrams.pyholds two example diagrams. Replace them with the real ones. Make sure that the repository ignores__pycache__/.
The package lives in the project, so the project can build its diagrams
without this skill. To change the style, edit diagram/style.py.
Never pass colors or fonts from a diagram function.
3. Workflow
- Collect the facts. Take every node, edge, and label from the source of truth: code, configuration, infrastructure as code, manifests, and pipeline definitions. Draw the current state, unless the page that shows the diagram says that it shows a plan. Never invent a component to fill the layout.
- Choose the type (§4). Read its section in references/diagram-types.md. Read references/drawing-api.md before the first diagram.
- State the plan in one short message: the type, the nodes, and what the budget (§6) cuts or splits. If the user can answer, let them redirect you. If not, continue and list your assumptions in the report.
- Write one function for each diagram in
diagrams.py, and list the pages inPAGES. First sketch the grid of columns and rows, then compute the coordinates. - Download the icons and build (§7). Fix every error that the build reports, then run the taste gate (§8).
- If
Dhas a README index, update it when you add, rename, or remove a diagram. Delete the HTML and PNG files that the build reports as stale. Commitdiagrams.py, the HTML pages, and the PNGs together. Never edit the generated files by hand.
What ships with it
11 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.
- references/diagram-types.md 4.6 KB
- references/drawing-api.md 8.9 KB
- scripts/diagram/__init__.py 216 B runs code
- scripts/diagram/build.py 5.8 KB runs code
- scripts/diagram/geometry.py 4.1 KB runs code
- scripts/diagram/icons.py 6.6 KB runs code
- scripts/diagram/model.py 25 KB runs code
- scripts/diagram/raster.py 4.7 KB runs code
- scripts/diagram/rules.py 9.9 KB runs code
- scripts/diagram/style.py 2.6 KB runs code
- scripts/diagrams.py 4.3 KB runs code
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.
- today First seen · 179 lines · 108 tokens per session scan A b8a91a8f329b
diagram-create is a skill published in the GitHub repository yarlson/yarstack (3 stars, last pushed yesterday), licensed MIT. It adds 108 tokens to every session and 2,367 once invoked, about $0.0004 per session on Opus 5.5. 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 2026-11-06.
diagram-create compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| diagram-createyarlson/yarstack | Skill | 3 | yesterday | MIT | A | 108 | Claude Code, Codex |
| capturedavesheffer/hunch | Skill | 15 | yesterday | none | A | 21 | Claude Code |
| healdavesheffer/hunch | Skill | 15 | yesterday | none | A | 21 | Claude Code |
| fixdavesheffer/hunch | Skill | 15 | yesterday | none | A | 17 | Claude Code |
| whydavesheffer/hunch | Skill | 15 | yesterday | none | A | 18 | Claude Code |
| fragiledavesheffer/hunch | Skill | 15 | yesterday | none | A | 15 | Claude Code |
| safe-changecz1993/vibe-coding-repository-standard | Skill | 21 | 2mo ago | Apache-2.0 | A | 54 | Codex |