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/claudiu-deving/workflow-atlas/agents-mdgit clone --depth 1 https://github.com/claudiu-deving/workflow-atlasWrote 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/claudiu-deving/workflow-atlas/agents-md)<a href="https://agentmods.dev/instructions/claudiu-deving/workflow-atlas/agents-md"><img src="https://agentmods.dev/badge/instructions/claudiu-deving/workflow-atlas/agents-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 5.1 | $0.02226 | $0.02226 |
| Opus 5 | $0.01113 | $0.01113 |
| Sonnet 5 | $0.00445 | $0.00445 |
| Haiku 4.5 | $0.00223 | $0.00223 |
Grade A, and why
workflow-atlas AGENTS.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 5d 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 — 136 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Agent setup
Setup for any AI assistant (over MCP). Zero dependencies — no npm install.
Run it
- Via an MCP client (e.g. Claude Code): a project
.mcp.jsonregisters the server; the client spawnsnode server/server.mjsover stdio and manages it. In Claude Code, reload the session so.mcp.jsonis read, then approveworkflow-atlaswhen prompted (approval is a security boundary — it can't be auto-granted). - Standalone:
npm start(ornode server/server.mjs) → http://localhost:5174/.
Requires Node ≥ 20. Override the port with PORT=…. If that port is taken by an
unrelated process the server steps to the next free one; if another atlas
instance holds it, this one reuses that UI and runs as an MCP/stdio worker.
Tests
The runtime ships zero dependencies; tests use dev-only deps (npm install first).
npm run test:unit— pure-model unit tests (shared/board.js: validation, the cycle guard, path helpers). Fast, no browser.npm run test:e2e— Playwright drives the real server + Chromium against an isolated seeded project (.atlas-e2e-home, wiped each run). Covers the infinite nested zoom (dive 25 levels with the cam-scale staying O(1), seamless re-root, auto-reroot/pop on continuous wheel), the breadcrumb navigator, focus-depth persistence across reload, and direct-manipulation editing (inline title, drag, create, delete, space-pan).npm testruns both.
The canvas exposes window.__atlasCanvas (e.g. getNav(), popFocus(),
focusPath(), _cam(), _settled()) as the test/automation handle.
Projects. Each session works on ONE project — by default the git repo root
(so opening repo acme, or any of its worktrees, authors the one acme project),
else the launch directory's name when it isn't a git repo, or set $WORKFLOW_ATLAS_PROJECT
to override. Data lives under $WORKFLOW_ATLAS_HOME (~/.workflow-atlas by default), so
parallel sessions on different projects stay isolated. Every tool below acts on this
session's project.
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.
- 5d ago First seen · 136 lines · 2,226 tokens per session scan A 35dccc68cfbc
workflow-atlas AGENTS.md is an instructions file published in the GitHub repository claudiu-deving/workflow-atlas (0 stars, last pushed 2mo ago), licensed MIT. It adds 2,226 tokens to every session, about $0.0111 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
rust-2026-template AGENTS.md
AGENTS.md instructions for d-oit/rust-2026-template, covering agent coding contract, quick reference, project structure, agent skills (.agents/skills/) and multi-agent support.
rust-2026-template CLAUDE.md
Claude Code instructions for d-oit/rust-2026-template, covering claude adapter, canonical project rules live in agents.md (max 200 loc), this file contains only tool-specific differences, do not duplicate repo-wide instructions here and tool-specific guidance.
plumb-line AGENTS.md
AGENTS.md instructions for slopstopper/plumb-line, covering agents.md — plumb-line, where things are written down, four things that are load-bearing, tracking (recursive-spine convention) and moments map.
runecho CLAUDE.md
Claude Code instructions for inth3shadows/runecho, covering runecho workflow, release tags and git hooks — there are five, from two different installers.
vision-memory-mcp CLAUDE.md
Claude Code instructions for putervision/vision-memory-mcp, covering visual memory (vision-memory-mcp), 1. mandatory workflow & priority, 2. tool reference summary (15 core mcp tools), state memory (state-memory-mcp) and 1. priority order.
commander AGENTS.md
AGENTS.md instructions for autohandai/commander, covering all agents development standards, 🎯 mandatory requirements for all agents, you write the tdd and then write the feature implementation, architecture pattern - strict compliance and modular structure (required).