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/dork-labs/dorkos/browsergit clone --depth 1 https://github.com/dork-labs/dorkosWhat 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 | $0.00016 | $0.01226 |
| Opus 5 | $0.00008 | $0.00613 |
| Sonnet 5 | $0.00003 | $0.00245 |
| Haiku 4.5 | $0.00002 | $0.00123 |
Grade A, and why
browser scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
1. **Capture** — navigate to the URL; take a snapshot; collect console messages and network requests. Check the dev server is up (`curl -s http://localhost:${DORKOS_PORT:-6242}/api/health`) and scan recent server errors: How it starts
The opening of the file, as written. The whole thing — 56 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Browser Debug
Debug and resolve an issue observed in the browser — visual (design-system violations, layout), interaction (broken buttons/forms), data (wrong/stale content), or responsive. Parse $ARGUMENTS: an optional --url <url> flag plus the issue description. If the URL is missing, default to the Vite dev client at http://localhost:6241 (the Express API is on 6242; /api is proxied). If the description is too vague to act on, ask one clarifying question — otherwise just start.
The loop
Work the issue as an evidence-driven loop. Never edit code you haven't read, and never claim a fix you haven't re-verified in the browser.
-
Capture — navigate to the URL; take a snapshot; collect console messages and network requests. Check the dev server is up (
curl -s http://localhost:${DORKOS_PORT:-6242}/api/health) and scan recent server errors:LOG="apps/server/.temp/.dork/logs/dorkos.log" # NDJSON; level >= 40 is warn/error tail -200 "$LOG" | python3 -c " import sys, json for line in sys.stdin: try: o = json.loads(line) if o.get('level', 0) >= 40: print(f\"[{o.get('time','')}] {o.get('tag','?')}: {o.get('msg','')}\"[:200]) except: pass " | tail -20 -
Classify — visual (no console errors, looks wrong), interaction (errors on action), data (API failures, stale cache), performance (slow, timing warnings), or responsive (breaks at breakpoints).
-
Diagnose — trace the observation to source code and confirm the cause there before changing anything:
- Client code lives in
apps/client/src/(FSD layers underapps/client/src/layers/). - Visual issues: load the
designing-frontendskill for the design-system audit criteria andstyling-with-tailwind-shadcnfor token/component specs — don't re-derive the design system from memory. Take targeted element screenshots to compare against spec. - Interaction issues: attempt the interaction via browser tools, capture new console errors, then trace the event handler in the component.
- Data issues: trace component → TanStack Query hook (
apps/client/src/layers/entities/*/model/) → Express route (apps/server/src/routes/) → service (apps/server/src/services/); check server logs for the failing request. For deep data-flow work, switch to/debug:api. - Responsive issues: resize to 375×667, 768×1024, and 1280×800 and snapshot each; check
sm:/md:/lg:class usage and 44px touch targets. - For library behavior questions, query context7 (
resolve-library-id→query-docs). For multi-file hunts, dispatch acode-searchagent; for gnarly type errors mid-fix,typescript-expert; for CSS/layout expertise,react-tanstack-expert.
- Client code lives in
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 · 56 lines · 16 tokens per session scan A 2112837d56b4
browser is a command published in the GitHub repository dork-labs/dorkos (9 stars, last pushed 2d ago), licensed MIT. It adds 16 tokens to every session and 1,226 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other commands, from other repositories
dark-factory-behavioral-e2e
Holdout-pattern E2E validator. Drives agent-browser against the running DynaChat app to verify that the PR's user-facing behavior actually matches what the linked issue asked for.
testarch-framework
Initialize production-ready test framework architecture (Playwright or Cypress) with fixtures, helpers, and configuration.
auto-browse
Auto-browse — learn, optimize, and graduate browser operations or web data-mining workflows.
connect
Connect to OpenChrome and confirm the MCP server is reachable. Runs a quick health check, reports the Chrome CDP status, and prints the tool surface summary.
check-access
Check whether a user or group can access a specific URL via ZIA policies.
fest-show
Show festival progression (in-progress tasks, roadmap, and dependency view).