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 commands/othmanadi/mddesign/composegit clone --depth 1 https://github.com/OthmanAdi/MDDesignWhat 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 | $0.00044 | $0.00877 |
| Opus 5 | $0.00022 | $0.00439 |
| Sonnet 5 | $0.00009 | $0.00175 |
| Haiku 4.5 | $0.00004 | $0.00088 |
Grade B, and why
compose scanned grade B with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Enumerates other installed skillsmediumAgent snooping
Other skills' SKILL.md files reveal prompts, capabilities and secrets that should be invisible to peers.
- `Bash: ls "$HOME/.claude/skills/code-memory-router/SKILL.md" 2>/dev/null || ls "$HOME/.agents/skills/code-memory-router/SKILL.md" 2>/dev/null || ls "$HOME/.claude/plugins/cache/*/skills/code-memory-router/SKILL.md" 2>/ How it starts
The opening of the file, as written. The whole thing — 67 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/mddesign:compose
Run me at the start of work on a new project. I verify that MDDesign's four arms are wired and tell you what to install if anything is missing. Read-only; I never modify the project.
What I check
For the current project directory (CWD), I produce a status table and print it:
MDDesign compose check — <project name>
────────────────────────────────────────────────────────────
Arm | Status | Note
------------------|---------|--------------------------------------------
Planning | <ok|x> | task_plan.md / findings.md / progress.md
Design (read) | <ok|x> | DESIGN.md at project root
Design (harvest) | <ok|x> | npx @google/design.md lint reachable
Memory router | <ok|x> | code-memory-router slash command registered
Scratch tier | <ok|x> | .agents/memory/scratch/ directory writable
Recommended next step: <one-line action>
How I check each arm
Planning
Globfortask_plan.mdin CWD.- If absent: note "missing. Use planning-with-files to initialize."
- If present but
findings.mdorprogress.mdmissing, note "partial. Re-init or create the missing files."
Design (read)
GlobforDESIGN.mdin CWD.- If present: ok.
- If absent AND the project has UI files (
*.tsx|*.vue|*.svelte|*.html|*.css): note "missing. Run/mddesign:harvestto generate one from your codebase." - If absent AND no UI files detected: note "not applicable (no UI files detected)."
Design (harvest toolchain)
Bash: npx --yes @google/design.md --version 2>/dev/null || echo "not-available"- If output contains
not-available: note "npx @google/design.md not reachable. Run/mddesign:setupto install, or ensure Node and npx are on PATH."
Memory router
- Look for
code-memory-routerSKILL.md in plugin cache or user skill dir. Bash: ls "$HOME/.claude/skills/code-memory-router/SKILL.md" 2>/dev/null || ls "$HOME/.agents/skills/code-memory-router/SKILL.md" 2>/dev/null || ls "$HOME/.claude/plugins/cache/*/skills/code-memory-router/SKILL.md" 2>/dev/null || echo "not-found"- If not found: note "code-memory-router not installed. /mddesign:memory recall will fall back to scratch only. Run
/mddesign:setup."
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.
- 2d ago First seen · 67 lines · 44 tokens per session scan B 65c34396a548
compose is a command published in the GitHub repository OthmanAdi/MDDesign (13 stars, last pushed 4mo ago), licensed Apache-2.0. It adds 44 tokens to every session and 877 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it B with 1 finding (enumerates other installed skills). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other commands, from other repositories
browser
Inspect and control the current AO session's target-isolated browser. The desktop app must be open. The agent and user share the same live page, cookies, navigation state, and WebContentsView; the runtime remains usable while the Browser panel is hidden. Tabs in this worker share an ephemeral browser profile, while…
preview
Open a URL or workspace file in the desktop browser panel for the current session, or start a deterministic session-owned dev server from an existing .ao/launch.json.
project
Manage projects: register repos, inspect, configure per-project settings, and remove.
session
Manage agent sessions: list, inspect, rename, kill, restore, clean up, and claim PRs.
spawn
Spawn a worker agent session in a registered project. The session runs the chosen agent in a fresh git worktree. Register the project first with ao project add.
import
Import reads the legacy Agent Orchestrator flat-file store (/.agent-orchestrator) read-only and ports its projects and per-project settings into the rewrite database. Legacy files are never modified, and a re-run skips rows that already exist, so it is safe to run more than once. The daemon must be stopped before…