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/ucsandman/dashclaw/dashclaw-qualitygit clone --depth 1 https://github.com/ucsandman/DashClawWrote 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/commands/ucsandman/dashclaw/dashclaw-quality)<a href="https://agentmods.dev/commands/ucsandman/dashclaw/dashclaw-quality"><img src="https://agentmods.dev/badge/commands/ucsandman/dashclaw/dashclaw-quality.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 | $0.00053 | $0.01236 |
| Opus 5 | $0.00026 | $0.00618 |
| Sonnet 5 | $0.00011 | $0.00247 |
| Haiku 4.5 | $0.00005 | $0.00124 |
Grade A, and why
dashclaw-quality 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 4d 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 — 80 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Goal: make the DashClaw site actually work, without me clicking through it
You are running a recurring find-and-fix quality loop on this repo. The goal is a clean app: every page renders, every code gate is green, and the fixes are in the working tree ready to ship. Drive it to that state autonomously — do not hand me a checklist of things for me to do.
This command explicitly authorizes you to call the Workflow tool (the
dashclaw-find-and-fix script in .claude/workflows/). That is the engine; you are the driver.
Stop condition
Loop until the workflow reports stillBroken: [] and a full gate run is green, OR until
two consecutive rounds land nothing new (then stop and report what's stuck — do not spin).
Hard cap: 3 rounds. If $ARGUMENTS names a focus area, scope triage/fixes to it. To go
deeper, include a token target like +500k in your message — the workflow's fan-out (finder
count and number of issues fixed) auto-scales to the turn's budget.total; it's read from the
harness, not passed as an arg.
Each round
-
Boot the browser layer (this session owns the dev server). Check
http://localhost:3000. If it isn't up, startnpm run devin the background and wait until it answers. The dev server must be started here, not inside the workflow — background processes started by a sub-agent don't reliably survive.- If a stale
.next/devcache makes every route 500 whilenpm run buildpasses, clear it and restart dev before trusting the smoke results.
- If a stale
-
Browser smoke with
frontend-verify. Invoke thefrontend-verifyskill across the routes (it's the token-cheap verifier — reads console errors + failed network requests first, writes full page state to disk; this is not the Chrome DevTools MCP, which stays banned). Collect its findings as a list of{source:"browser", severity, summary, location, evidence}. -
Run the workflow. Call
Workflowwithname: "dashclaw-find-and-fix"andargs: { baseUrl: "http://localhost:3000", browserFindings: [...the list from step 2...] }. It will: discover routes → HTTP-smoke + run lint/vitest/build/contract gates in parallel → triage everything (your browser findings included) into atomic issues → fix the top ones in isolated worktrees in parallel → integrate the green diffs sequentially onto the working tree.
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.
- 4d ago First seen · 80 lines · 53 tokens per session scan A ac37af3ab352
dashclaw-quality is a command published in the GitHub repository ucsandman/DashClaw (296 stars, last pushed 5d ago), licensed MIT. It adds 53 tokens to every session and 1,236 once invoked, about $0.0003 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 2026-08-30.
Other commands, from other repositories
hep-browser
Use the Agentlas browser hardpoint for browser-required work.
agentlas-browser
Use the Agentlas browser hardpoint for browser-required work.
agentlas-login
Sign this machine into Agentlas (opens the browser sign-in window).
code-review
Provide a code review for the given pull request.
echo
This is a placeholder command definition for plugin framework tests.
auto-verify
프론트엔드 UX 검증 — Playwright 기반 비주얼 검증을 실행합니다.