Borrowing it
Nothing to install: this file belongs to rakaarwaky/qwen-web-arwaky. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/rakaarwaky/qwen-web-arwaky/main/.agents/skills/qwen-web/SKILL.mdgit clone --depth 1 https://github.com/rakaarwaky/qwen-web-arwakyWrote 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/skills/rakaarwaky/qwen-web-arwaky/qwen-web)<a href="https://agentmods.dev/skills/rakaarwaky/qwen-web-arwaky/qwen-web"><img src="https://agentmods.dev/badge/skills/rakaarwaky/qwen-web-arwaky/qwen-web/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/rakaarwaky/qwen-web-arwaky/qwen-web"><img src="https://agentmods.dev/badge/skills/rakaarwaky/qwen-web-arwaky/qwen-web.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 2 findings, up to high
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- high Rogue Agent · line 146 Skill modifies its own code, configuration, or behavior at runtime. Self-modification enables an agent to escalate privileges, disable safety constraints, or install persistent backdoors.Fix: Prevent the skill from modifying its own code, SKILL.md, or configuration files. Treat skill files as read-only at runtime.
- medium Output Handling · line 228 Output size or generation rate is not bounded. Unbounded output enables denial-of-service through resource exhaustion, log flooding, or context-window stuffing.Fix: Set explicit limits on output length, generation count, and rate. Use max_tokens and truncation to prevent unbounded output.
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 | $0.00121 | $0.06564 |
| Opus 5 | $0.00060 | $0.03282 |
| Sonnet 5 | $0.00024 | $0.01313 |
| Haiku 4.5 | $0.00012 | $0.00656 |
Grade A, and why
qwen-web 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 yesterday.
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 — 509 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Qwen Web Automation — Master-Class Agent Harness
Purpose. This skill turns
qwen-web-arwakyinto a deterministic, production-grade LLM backend for autonomous AI agents. It covers every MCP tool, every CLI subcommand, prompt-engineering contracts for complete un-truncated output, multi-role engineering workflows, and a full resilience decision tree.Prime directive for agents: treat Qwen3.8-Max as your deep-reasoning engine. Structure every request with an explicit Output Contract, pick the correct timeout band, and always verify the output envelope before consuming results.
1. Core Capabilities & Architecture Summary
1.1 What makes this engine different
| Capability | Implementation detail |
|---|---|
| Qwen3.8-Max by default | The pipeline forces and verifies the hardcoded default model (Qwen3.8-Max) on every session before dispatch. The agent never picks a model manually; a ModelSwitchError aborts the run if the model cannot be confirmed active. |
| Zero API key | Authentication is a real Chromium browser session (persistent cookies in qwen_session/). No tokens, no billing, no quota. |
| Zero-truncation extraction (Tier-1 React Fiber) | Response capture walks the React Fiber tree (__reactFiber → memoizedProps.content, up to 30 levels) to recover 100% of raw Markdown, including code blocks that Monaco Editor virtualization would clip in the visible DOM. |
| Tier-2 DOM Tree Walker fallback | A block-aware tree walker (preserves PRE/code formatting, strips UI chrome, buttons, copy widgets) guarantees extraction even when Fiber props are absent. |
| 30s cloud reload sync | While waiting for a terminal event, the page is reloaded every 30 seconds to re-sync Qwen Cloud streaming state — long generations survive network drops and tab throttling. |
| Event-driven long-run monitor | The stream monitor has no elapsed-time response cutoff. It waits for terminal generation state and keeps recovery polling alive for long-running jobs. |
| Terminal-event completion | A response is accepted only after stable text and an explicit generation-complete signal; stability alone can never terminate an unfinished response. |
| Self-healing browser sessions | Automatic stale SingletonLock cleanup, session directory permission repair (0700), 3-attempt launch retry with backoff. |
| Multi-strategy input injection | Playwright fill → React native value-setter (with _valueTracker reset) → type() keystroke fallback. |
| Parse-gated dispatch | The send button is held until document parsing is positively verified (network files/parse 200 + DOM spinner/toast clearance). Prompts are never dispatched onto half-parsed attachments. |
| Atomic, traceable output | Outputs are written atomically with a METADATA TRACEABILITY header and a .meta.json sidecar (run_id, duration_sec, input_chars, output_chars). |
| Full observability | Structured JSONL logs (app.jsonl), status.json for monitoring, optional Sentry + OpenTelemetry tracing, lifecycle event stream with strict predecessor gating. |
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.
- yesterday Changed 2f013ce694f2
- 2d ago Changed · +36 lines 39cd5621348d
- 11d ago First seen · 473 lines · 121 tokens per session scan A dcebb5043c22
qwen-web is a skill published in the GitHub repository rakaarwaky/qwen-web-arwaky (13 stars, last pushed yesterday), licensed MIT. It adds 121 tokens to every session and 6,564 once invoked, about $0.0006 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 skills, from other repositories
debug-optimize-lcp
Guides debugging and optimizing Largest Contentful Paint (LCP) using Chrome DevTools MCP tools. Use this skill whenever the user asks about LCP performance, slow page loads, Core Web Vitals optimization, or wants to understand why their page's main content takes too long to appear. Also use when the user mentions…
opencli-sitemap-author
Use when creating or maintaining OpenCLI site sitemaps: agent-facing navigation, page-state, action, workflow, API-reference, pitfall, and fallback knowledge for a website. Use after browser exploration discovers durable site context, when a sitemap is stale, or when promoting local site knowledge into the repo.
interactive-login
How to complete browser/interactive logins (aws / gh / glab / gcloud). The platform backgrounds the login poller so it survives the human's browser round-trip — and when that does NOT work.
pinchtab-mcp
Use this skill when a task requires browser automation through PinchTab's MCP server connected to a remote browser instance. Covers navigation, element interaction, data extraction, form filling, multi-step flows, and session management via MCP tools.
azure-messaging-webpubsub-java
Build real-time web applications with Azure Web PubSub SDK for Java. Use when implementing WebSocket-based messaging, live updates, chat applications, or server-to-client push notifications.
google-safe-browsing
Prevent and fix Google Safe Browsing "Dangerous site" flags. Use when launching a public web app, buying/picking a domain, building a login or signup page, or when any site shows a red "Dangerous site" / "Deceptive site" warning in Chrome, Brave, Safari, Firefox, or Edge. Triggers on "dangerous site", "deceptive…