Skip to content

chore: mark v0.2.0 #3

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
b1zya merged 1 commit into from
Sep 5, 2026
Merged

chore: mark v0.2.0 #3

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
30 changes: 15 additions & 15 deletions CHANGELOG.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,7 +4,22 @@ All notable changes to this project are documented here. The format follows [Kee

## [Unreleased]

Nothing yet.

## [0.2.0] - 2026-09-05

First release as patchright-cli: the playwright-cli fork rebuilt as an anti-detection CLI for AI agents.

### Added
- Own TypeScript client (forked from playwright-cli's cli-client) driving the untouched `patchright-core` daemon; own state root, session registry, sockets and skill installer, so a playwright-cli on the same machine is never touched.
- Stealth defaults for `open`: real Google Chrome (Edge fallback), headed and maximized, persistent per-session profile, `viewport: null`, real OS color-scheme and motion preferences.
- `--proxy` with geoip: timezone, `navigator.languages`/`Accept-Language` and geolocation derived from the exit IP; WebRTC restricted to proxied routes; credentials kept off the command line and out of `config-print`.
- Identity file per profile, `identity`, `doctor` and `selftest` (38 main-world checks, `--online` detector tour).
- Leak warnings for every deviation (`--headless`, `--device`, manual geo, routes, overlays, ...), two of them unsilenceable.
- Guards: `run-code` safety scan, `eval --main-world`, `resize` moving the real window, `console` explained instead of re-enabling `Runtime.enable`, Firefox/WebKit and extension attach refused.
- Optional humanized input (`--humanize`): Bézier pointer paths with deceleration, click hold times, per-character typing.
- Layered configuration (`~/.patchright-cli/config.json`, `.playwright/patchright-cli.config.json`, `sessions.<name>`), generated daemon config, `check-drift` and `roll` tooling, CI on Windows/macOS/Linux, upstream watch.
- Agent skill rewritten around the anti-detection contract.
- `env`: how to run commands on this machine — shell quoting rules (PowerShell, cmd.exe, Git Bash, POSIX), agent host and sandbox markers, writable state root, console encoding, display, host locale/timezone, default browser; `--probe` checks network access.
- `batch`: run a whole flow (open … close) from a script in one invocation, for sandboxes that kill background processes between tool calls and for permission modes that confirm every command.
- Notice `msys-path-conversion` when Git Bash turned a leading-slash argument into a Windows path.
Expand Down Expand Up @@ -35,21 +50,6 @@ All notable changes to this project are documented here. The format follows [Kee
- `eval --main-world` with a bare expression (`document.title`) ran in the daemon instead of the page.
- `console` explanation, warning and docs now say precisely what is missing (`console.*` calls and uncaught exceptions, which need `Runtime.enable`) and what is still listed (browser-level log entries: failed requests, CSP violations), instead of "the Console API is disabled".

## [0.2.0] - 2026-09-05

First release as patchright-cli: the playwright-cli fork rebuilt as an anti-detection CLI for AI agents.

### Added
- Own TypeScript client (forked from playwright-cli's cli-client) driving the untouched `patchright-core` daemon; own state root, session registry, sockets and skill installer, so a playwright-cli on the same machine is never touched.
- Stealth defaults for `open`: real Google Chrome (Edge fallback), headed and maximized, persistent per-session profile, `viewport: null`, real OS color-scheme and motion preferences.
- `--proxy` with geoip: timezone, `navigator.languages`/`Accept-Language` and geolocation derived from the exit IP; WebRTC restricted to proxied routes; credentials kept off the command line and out of `config-print`.
- Identity file per profile, `identity`, `doctor` and `selftest` (38 main-world checks, `--online` detector tour).
- Leak warnings for every deviation (`--headless`, `--device`, manual geo, routes, overlays, ...), two of them unsilenceable.
- Guards: `run-code` safety scan, `eval --main-world`, `resize` moving the real window, `console` explained instead of re-enabling `Runtime.enable`, Firefox/WebKit and extension attach refused.
- Optional humanized input (`--humanize`): Bézier pointer paths with deceleration, click hold times, per-character typing.
- Layered configuration (`~/.patchright-cli/config.json`, `.playwright/patchright-cli.config.json`, `sessions.<name>`), generated daemon config, `check-drift` and `roll` tooling, CI on Windows/macOS/Linux, upstream watch.
- Agent skill rewritten around the anti-detection contract.

### Removed
- npm publishing, Microsoft-internal pipelines, Playwright test-generation references, `attach --extension`.

Expand Down
Loading