Measured result, selection split
Taken from the research repo's README. Held-out test pending.
Qwen3.8-27B, driven by the pi coding agent, 8 selection tasks with 6 runs each (48 rollouts per skill). The held-out test is pending: those runs were still in progress when this was written, so there is no unseen-task result yet.
| Skill | Edits | Correct | Score | Tokens per task | Turns | Decision |
|---|---|---|---|---|---|---|
| No skill (24 rollouts) | 21/24 | 0.7887 | 68.7k | 13.9 | ||
s1 (Opus-trained) | 47/48 | 0.9177 | 42.3k | 9.2 | start | |
q1 | 4 | 43/48 | 0.8367 | 42.0k | 8.3 | rejected |
q2 | q1 with one edit swapped | 48/48 | 0.9454 | 36.4k | 8.5 | accepted |
q3 | 4 more | 47/48 | 0.9215 | 39.4k | 8.1 | rejected |
q4 | 2 of the q3 edits | 44/48 | 0.8542 | 47.0k | 9.1 | see note |
Swipe the table to see more.
q4 note: three of its four failures were 15-minute timeouts on the 50-page catalogue crawl, after the practice site began answering 403 under four concurrent crawls. That is the benchmark throttling, not the skill.
The skill file is q2 plus the same hand-added Setup section as Agent Browse.
What it does
A variant of Agent Browse for Qwen3.8-27B and similar local models that tend to guess selectors and URLs. It teaches an assistant to drive a real browser with the agent-browser command line tool: clicking, filling forms, logging in, hover menus, uploads, JavaScript dialogs, paginated data.
It was trained the same way as Agent Browse: scored tasks, and only edits that measurably helped were kept. It starts from the skill trained for Claude Opus and adds what the Qwen runs showed: never invent a URL or selector you have not seen, use the CLI's own selector syntax, pass longer JavaScript on stdin, check a list against the page's own total before calling it complete, and do not close the browser early.
The test harness, every candidate version (accepted and rejected) and the run data are in the research repo: https://github.com/proskillpacks/skillopt-agent-browse (Part 2 of the write-up).
The evidence
OPTIMISED badge. Improved step by step against a set of test cases, with every score on file.Improved over 4 gated rounds: 48 of 48 on the selection tasks.
Trained and scored with Qwen3.8-27B on a local server, driven by the pi coding agent. On the 8 selection tasks (6 runs each) it got 48 of 48 correct at 36.4k tokens per task, against 21 of 24 at 68.7k with no skill and 47 of 48 at 42.3k with the Claude-trained Agent Browse. These are selection-split figures, the split the gate used to pick it. The held-out test is pending. Three of four tuning attempts were rejected by the gate.
More on how it was tested
The held-out test for this version is not finished. We did not score it with any other model.
Limits
- Held-out test pending: the figures are on the selection split, which the gate used to choose this skill.
- The cross-check on Claude Opus was not run.
- One local model, one quantisation, one harness (pi). Other models are untested.
- The selection runs were measured under four-way concurrency and site throttling.
- The Setup section of SKILL.md was added by hand and did not go through the validation gate.
Questions
Questions or a problem?
Which assistants does it work with?
Tested with Qwen3.8-27B on a local OpenAI-compatible server, driven by the pi coding agent, and the agent-browser CLI (https://github.com/vercel-labs/agent-browser). The format is the open SKILL.md format; other assistants and models are untested. You need an assistant that can run shell commands.
For technical users
Using it with an AI assistant
- Install the agent-browser CLI from https://github.com/vercel-labs/agent-browser and check that
agent-browser --helpruns. - Copy the
agent-browse-qwenfolder from GitHub (https://github.com/proskillpacks/skills/tree/main/skills/developers/agent-browse-qwen) into.agents/skills/in your project, or into your tool's own skills folder. - Ask for a browser task in plain words, for example: "Open the page, find the price table and give me the rows." The skill loads by itself. There is no paste-in prompt for this one, because it needs a shell.
Jobs that go with this
- Find accessibility problems on a web page (Free): A page address in, a ranked list of problems with a fix for each.
- Find code changes that have no tests (Free): A diff and its tests in, what is not covered and test cases to add.
- Write a changelog from your commits (Free): A list of commits in, release notes people can read.