mermaid-diagrams
Writes Mermaid diagrams as plain-text fenced code blocks, entirely locally, and never sends diagram source to an online renderer, editor, or API.
Skill by twistingmercury
Chalk Labs
Need help deploying mermaid-diagrams?
Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/twistingmercury-mermaid-diagrams
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: documents, design, backend
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md and code. Checked Nov 1, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/mermaid-diagrams
What's inside
- Skill name
mermaid-diagrams- Description
- Writes Mermaid diagrams as plain-text fenced code blocks, entirely locally, and never sends diagram source to an online renderer, editor, or API. Use when a document, design, or README needs a flowchart, sequence, ER, C4, state, class, or Gantt diagram, or when an existing Mermaid block has a syntax error.
- License
- MIT
- Instructions
- 100 lines in SKILL.md · view on GitHub
Files (3)
- SKILL.md
- local-validation.md · references/
- syntax-pitfalls.md · references/
SKILL.md (first lines)
# Mermaid Diagrams Skill
Write Mermaid diagrams as fenced code blocks in Markdown documents. Diagrams stay local, never sent
to an online service.
## Local only
Diagrams live in the document as fenced `mermaid` blocks. Never paste or send diagram source to
mermaid.live, mermaid.ink, Kroki, or any other online renderer, editor, or API, and never generate
diagram image files through an online service. Do not install tools or use `npx`.
## Colors that work in light and dark themes
Never set `theme` (not in frontmatter config, not via `%%{init}%%`); renderers such as GitHub
follow the reader's theme for diagrams that do not choose one. Unstyled parts (edges, labels, subgraph boxes) adapt by
themselves. To give a node meaning-carrying color, use ONLY this palette and always set fill,
stroke, and color together in a `classDef`:
- stroke `#6e7781`, text `#1f2328` for every class
- blue `#9ec5f4`, green `#9fd8a8`, amber `#f2cc80`, red `#f2a9a0`, purple `#c9b2ee`, gray `#c4cbd3`
Measured contrast (WCAG): text on every fill >= 8.2:1; stroke vs white canvas 4.55:1 and vs GitHub
dark canvas #0d1117 4.16:1. Color encodes meaning only (e.g. red = failure path), never decoration.
No pure white or black fills.
## Right angles
Mermaid 12 defaults to the ELK layout, which routes edges orthogonally with rounded corners: leave
the layout and curve defaults alone. Do not set `curve` (tested with mmdc 12.0.0: `curve: step` on
Dagre produces more crossings and overlapping edges). Never use curvy options (`basis`, `natural`).
Sequence diagrams are already rectilinear. These results come from mmdc 12.0.0 only; the Mermaid
version GitHub bundles may differ and was not tested.
## Few crossing lines
Declare nodes in layer order (top layer first); keep edges between adjacent layers; group related
nodes in `subgraph`; keep back-edges rare. If crossings cannot be avoided, split the diagram into
two. After rendering locally (see references), look at the picture and restructure if edges cross
avoidably.
## Choose the diagram type
| Type | Purpose |
| --- | --- |
| `flowchart` / `graph` | Processes, workflows, and architecture |
| `sequenceDiagram` | Interactions and flows over time |
| `erDiagram` | Data models and relationships |
| `C4Context` or `C4Container` | System context and containers |
| `stateDiagram-v2` | Lifecycles and state machines |
| `classDiagram` | Type structure and inheritance |
| `gantt` | Schedules and timelines |
## Workflow
1. Pick the diagram type (table above).
2. Write the block with a title in frontmatter (`title:`), labeled edges, and the palette rules.
3. Run the optional local check (references/local-validation.md).
4. Fix and re-run until clean.
## Optional local validation
If `mmdc` (mermaid-cli) is available locally, render to a temporary directory (never into the
repo), once light and once dark, then look at both images for legibility and avoidable crossings.
If `mmdc` is not available, apply the syntax pitfalls checklist by reading instead. See
references/local-validation.md for exact commands and where to look.
## Example: system context diagram
```mermaid
---
title: Payment Processing System Context
---
flowchart TD
subgraph boundary["Payment System"]
gateway["Payment Gateway"]:::blue
ledger["Ledger DB"]:::green
notify["Notifier"]:::amber
end
user["User"]
bank["External Bank"]
user -->|initiate| gateway
gateway -->|query| ledger
gateway -->|send| notify
gateway -->|transfer| bank
notify -->|confirm| user
classDef blue fill:#9ec5f4,stroke:#6e7781,color:#1f2328
classDef green fill:#9fd8a8,stroke:#6e7781,color:#1f2328
classDef amber fill:#f2cc80,stroke:#6e7781,color:#1f2328
```
## References
Read these when you need help or are debugging a diagram:
- **references/syntax-pitfalls.md** — read when a diagram fails to render or produces a misleading
picture; troubleshoot with concrete bad/fixed pairs
- **references/local-validation.md** — read when you want to check a diagram locally before
committing; covers mmdc options, light/dark rendering, and what to look for in the imagesFrom twistingmercury/agents-skills, MIT.
More from twistingmercury/agents-skills
- 0.570score
python-uv-starter
twistingmercury
Scaffolds an empty directory as a Python CLI project with uv, src/ layout, pytest, ruff, a Makefile, a Docker-first wheel build, and GitHub Actions CI.
SkillAFreshness: AClaude Code +9 - 0.570score
rlm
twistingmercury
Runs a Recursive Language Model-style loop over a large file or document directory using a persistent local REPL: load the context once, then search, chunk…
SkillAFreshness: AClaude Code +9 - 0.570score
docker-first-ci
twistingmercury
Implements and hardens Docker-first build, test, and release pipelines, where one build script runs the quality gates, the build, and the e2e tests inside…
SkillAFreshness: AClaude Code +9 - 0.565score
shell-script
twistingmercury
Create or modify production-grade shell scripts with automatic BATS test coverage and a bounded verification loop.
SkillAFreshness: AClaude Code +9 - 0.565score
capture-requirements
twistingmercury
Elicit, document, and refine software project requirements for handoff to architecture and design agents.
SkillAFreshness: AClaude Code +9 - 0.565score
check-push-readiness
twistingmercury
Audits, read-only, the exact unpushed commits before a push: resolves the live remote SHA, checks fast-forward, scans the diff for secrets and debug code…
SkillAFreshness: AClaude Code +9 - 0.565score
design-docs-writer
twistingmercury
Creates and updates architecture documents under docs/design/ from standardized templates, using snake_case filenames and frontmatter versioning, always…
SkillAFreshness: AClaude Code +9 - 0.565score
readme-writer
twistingmercury
Creates or updates a concise root README.md from a required template baseline.
SkillAFreshness: AClaude Code +9 - 0.561score
code-review
twistingmercury
Conduct evidence-based code reviews of files, diffs, pull requests, or working changes using independent general, architectural, and language perspectives.
SkillAFreshness: AClaude Code +9 - 0.559score
prime
twistingmercury
Surveys a repository to build working context by reading its README and listing tracked files.
SkillAFreshness: AClaude Code +9
Signals
- Stars
- 0
- Forks
- 0
- Open issues
- 0
- Last commit (this folder)
- Oct 5, 2026 (1 month ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- Repository
- github.com/twistingmercury/agents-skills
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from twistingmercury
- 0.570score
python-uv-starter
twistingmercury
Scaffolds an empty directory as a Python CLI project with uv, src/ layout, pytest, ruff, a Makefile, a Docker-first wheel build, and GitHub Actions CI.
SkillAFreshness: AClaude Code +9 - 0.570score
rlm
twistingmercury
Runs a Recursive Language Model-style loop over a large file or document directory using a persistent local REPL: load the context once, then search, chunk…
SkillAFreshness: AClaude Code +9 - 0.570score
docker-first-ci
twistingmercury
Implements and hardens Docker-first build, test, and release pipelines, where one build script runs the quality gates, the build, and the e2e tests inside…
SkillAFreshness: AClaude Code +9 - 0.565score
shell-script
twistingmercury
Create or modify production-grade shell scripts with automatic BATS test coverage and a bounded verification loop.
SkillAFreshness: AClaude Code +9 - 0.565score
capture-requirements
twistingmercury
Elicit, document, and refine software project requirements for handoff to architecture and design agents.
SkillAFreshness: AClaude Code +9 - 0.565score
check-push-readiness
twistingmercury
Audits, read-only, the exact unpushed commits before a push: resolves the live remote SHA, checks fast-forward, scans the diff for secrets and debug code…
SkillAFreshness: AClaude Code +9
Done for you
Rather not set it up yourself?
Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.
See how we build AI agents →Read the case studies
Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.
- launches shipped
- 50+
- launches shipped
- media placements
- 500+
- media placements
- founders on every account
- 2
- founders on every account
Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block
Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.
Get mermaid-diagrams running in your business
Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.
Cite this page
mermaid-diagrams skill. Agent Mart. https://agentmart.dev/skills/twistingmercury-mermaid-diagrams (last updated 2026-10-24)