ONNX Runtime Shape Inference Safety Audit
microsoft/onnxruntime
Finds and fixes out-of-range output writes in ONNX Runtime operator shape-inference functions where a getNumOutputs guard admits too few outputs.
Launch stealth Chromium with C++ fingerprint patches for anti-bot bypass.
$ npx skills add uphiago/recon-skills --skill stealth-browser-launch -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install uphiago/recon-skills stealth-browser-launch --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/uphiago/recon-skills.git skills-src && mkdir -p .claude/skills && cp -r skills-src/recon/stealth-browser-launch .claude/skills/stealth-browser-launch && rm -rf skills-srcUse ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.
Claude Code skills documentation · loads skills from .claude/skills/
Install the "stealth-browser-launch" agent skill from https://github.com/uphiago/recon-skills/tree/main/recon/stealth-browser-launch into .claude/skills/stealth-browser-launch/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "stealth-browser-launch", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/uphiago/recon-skills/tree/main/recon/stealth-browser-launchType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add uphiago/recon-skills --skill stealth-browser-launch -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install uphiago/recon-skills stealth-browser-launch --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/uphiago/recon-skills.git skills-src && mkdir -p .agents/skills && cp -r skills-src/recon/stealth-browser-launch .agents/skills/stealth-browser-launch && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "stealth-browser-launch" agent skill from https://github.com/uphiago/recon-skills/tree/main/recon/stealth-browser-launch into .agents/skills/stealth-browser-launch/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "stealth-browser-launch", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add uphiago/recon-skills --skill stealth-browser-launch -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install uphiago/recon-skills stealth-browser-launch --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/uphiago/recon-skills.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/recon/stealth-browser-launch .cursor/skills/stealth-browser-launch && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "stealth-browser-launch" agent skill from https://github.com/uphiago/recon-skills/tree/main/recon/stealth-browser-launch into .cursor/skills/stealth-browser-launch/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "stealth-browser-launch", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/uphiago/recon-skills.git --path recon/stealth-browser-launch--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add uphiago/recon-skills --skill stealth-browser-launch -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install uphiago/recon-skills stealth-browser-launch --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/uphiago/recon-skills.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/recon/stealth-browser-launch .gemini/skills/stealth-browser-launch && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "stealth-browser-launch" agent skill from https://github.com/uphiago/recon-skills/tree/main/recon/stealth-browser-launch into .gemini/skills/stealth-browser-launch/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "stealth-browser-launch", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install uphiago/recon-skills stealth-browser-launchInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add uphiago/recon-skills --skill stealth-browser-launch -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/uphiago/recon-skills.git skills-src && mkdir -p .github/skills && cp -r skills-src/recon/stealth-browser-launch .github/skills/stealth-browser-launch && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "stealth-browser-launch" agent skill from https://github.com/uphiago/recon-skills/tree/main/recon/stealth-browser-launch into .github/skills/stealth-browser-launch/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "stealth-browser-launch", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add uphiago/recon-skills --skill stealth-browser-launch -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install uphiago/recon-skills stealth-browser-launch --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/uphiago/recon-skills.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/recon/stealth-browser-launch .opencode/skills/stealth-browser-launch && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "stealth-browser-launch" agent skill from https://github.com/uphiago/recon-skills/tree/main/recon/stealth-browser-launch into .opencode/skills/stealth-browser-launch/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "stealth-browser-launch", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
stealth-browser-launchLaunch stealth Chromium with C++ fingerprint patches for anti-bot bypass.
Stealth Browser Launch is an agent skill from uphiago/recon-skills. Launch stealth Chromium with C++ fingerprint patches for anti-bot bypass.
Its SKILL.md is about 2.2k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts. Compatibility notes: Requires curl, httpx, nuclei, python3
It sits in Security. It works with C++ and Playwright. The repository describes itself as: Recon & pentest skill pack. CORS, XSS, SQLi, SSRF, RCE, WordPress, MCP, cloud, subdomain takeover, and more. Field-tested. MIT. Full write-up at hiago.sh. The licence is MIT.
7 steps, taken from the step headings in SKILL.md.
Read from SKILL.md and the folder at commit 1260244. It shows what the files ask for, not the result of running them.
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Shell commands in SKILL.md call:
pippython3dockeraptplaywrightFrom the folder's file list and the shell code blocks in SKILL.md.
Hosts in commands or code, which the agent is likely to contact:
browserscan.netdemo.fingerprint.comantcpt.comdeviceandbrowserinfo.comFrom URLs in SKILL.md, links to its own repository left out.
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Requires curl, httpx, nuclei, python3
From compatibility in the SKILL.md frontmatter.
Stealth Browser Launch loads about 2.2k tokens when it runs. Until then it costs about 24 tokens; SKILL.md has 540 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.
The full file from uphiago/recon-skills at commit 1260244, republished under its MIT licence (© uphiago). 540 words, ~2,155 tokens.
.claude/skills/stealth-browser-launch/SKILL.md (or your agent's skills folder).Launch a patched Chromium binary with C++ source-level fingerprint modifications that bypass anti-bot detection. The binary spoofs canvas, WebGL, audio, GPU, screen, WebRTC, network timing, and automation signals at the binary level — not via JavaScript injection or config patches that break with Chrome updates. Passes Cloudflare Turnstile, reCAPTCHA v3 (0.9 score), FingerprintJS, BrowserScan, and 30+ detection sites.
terminal with python3 and pip.playwright installed: pip install playwright && playwright install-deps chromium.cloakbrowser[geoip] for automatic timezone/locale resolution from proxy exit IP.pip install cloakbrowser
python3 -c "
from cloakbrowser import launch
browser = launch()
page = browser.new_page()
page.goto('https://target.com')
print(page.title())
browser.close()
"The binary auto-generates a random fingerprint seed per launch. No flags needed for basic stealth:
from cloakbrowser import launch
browser = launch(
headless=True,
proxy="http://user:pass@residential-proxy:port",
geoip=True, # auto-detect timezone/locale from proxy exit IP
)
page = browser.new_page()
page.goto("https://target.com")
# Standard Playwright API from here
page.locator("input[name='username']").fill("test")
page.locator("button[type='submit']").click()
browser.close()Use a fixed fingerprint seed when revisiting the same target to appear as a returning visitor:
browser = launch(
proxy="http://user:pass@residential-proxy:port",
geoip=True,
args=["--fingerprint=42069"], # fixed seed = same fingerprint every launch
)Some sites detect headless even with C++ patches. Run headed with a virtual display:
# Start virtual display
Xvfb :99 -screen 0 1920x1080x24 &
export DISPLAY=:99browser = launch(
headless=False, # real rendering
proxy="http://residential-proxy:port",
geoip=True,
humanize=True, # human-like mouse/keyboard/scroll
)Override specific fingerprint values to match a target environment:
browser = launch(stealth_args=False, args=[
"--fingerprint=42069",
"--fingerprint-platform=windows",
"--fingerprint-gpu-vendor=NVIDIA Corporation",
"--fingerprint-gpu-renderer=NVIDIA GeForce RTX 3060/PCIe/SSE2",
"--fingerprint-hardware-concurrency=16",
"--fingerprint-device-memory=16",
"--fingerprint-screen-width=2560",
"--fingerprint-screen-height=1440",
"--fingerprint-timezone=America/Sao_Paulo",
"--fingerprint-locale=pt-BR",
"--fingerprint-webrtc-ip=auto",
"--fingerprint-noise=false", # disable noise for FingerprintJS ML bypass
])Maintain cookies and localStorage across sessions to bypass "first-visit" challenges:
from cloakbrowser import launch_persistent_context
ctx = launch_persistent_context(
"./target-profile",
headless=False,
proxy="http://residential-proxy:port",
geoip=True,
)
page = ctx.new_page()
page.goto("https://target.com")
# Session persists across restarts
ctx.close()Run multiple browser identities from a single container using cloakserve:
docker run -d --name cloak -p 127.0.0.1:9222:9222 cloakhq/cloakbrowser cloakservefrom playwright.sync_api import sync_playwright
pw = sync_playwright().start()
# Each unique fingerprint seed spawns separate Chrome process with independent identity
b1 = pw.chromium.connect_over_cdp("http://localhost:9222?fingerprint=11111")
b2 = pw.chromium.connect_over_cdp("http://localhost:9222?fingerprint=22222")
# Full identity control via query params
b3 = pw.chromium.connect_over_cdp(
"http://localhost:9222?fingerprint=33333"
"&timezone=America/New_York&locale=en-US&platform=windows"
"&hardware-concurrency=4&device-memory=8"
)
# Each browser has independent cookies, localStorage, canvas noiseFor aggressive sites (Kasada, Akamai) that check canvas emoji rendering:
apt install -y fonts-noto-color-emoji fonts-freefont-ttf fonts-unifont \
fonts-ipafont-gothic fonts-wqy-zenhei fonts-tlwg-loma-otfFor CreepJS font enumeration evasion, install Windows fonts:
# Copy from Windows machine: C:\Windows\Fonts\
mkdir -p ~/.local/share/fonts/windows
cp /path/to/windows/fonts/SegoeUI*.ttf ~/.local/share/fonts/windows/
fc-cache -f
# Then launch with:
browser = launch(args=["--fingerprint-fonts-dir=/home/user/.local/share/fonts/windows"])| Flag | Default | What it controls |
|---|---|---|
--fingerprint=SEED | Random 5-digit | Master seed for canvas/WebGL/audio/fonts |
--fingerprint-platform | windows/macos | navigator.platform, UA OS, GPU pool |
--fingerprint-gpu-vendor | Auto | WebGL UNMASKED_VENDOR_WEBGL |
--fingerprint-gpu-renderer | Auto | WebGL UNMASKED_RENDERER_WEBGL |
--fingerprint-hardware-concurrency | 8 | navigator.hardwareConcurrency |
--fingerprint-device-memory | 8 | navigator.deviceMemory |
--fingerprint-screen-width | 1920/1440 | Screen width |
--fingerprint-screen-height | 1080/900 | Screen height |
--fingerprint-timezone | — | IANA timezone |
--fingerprint-locale | — | BCP 47 locale |
--fingerprint-webrtc-ip | — | WebRTC ICE IP (auto for proxy exit IP) |
--fingerprint-noise=false | true | Disable canvas/WebGL/audio noise |
--fingerprint-fonts-dir | — | Target platform fonts path |
--fingerprint-windows-font-metrics | — | Align font metrics to Windows (v148+) |
--fingerprint-storage-quota | Auto | Storage quota in MB |
--fingerprint-taskbar-height | 48/95/0 | Taskbar height spoofing |
page.wait_for_timeout() leaks CDP traffic that reCAPTCHA detects. Use time.sleep() instead.--fingerprint-noise=false can cause ML-based detection on FingerprintJS. Only disable noise when specifically blocked by it.headless=False) for maximum stealth on aggressive sites.browser_version= if you need reproducibility.https://browserscan.net — all 4 bot checks should show NORMAL.https://demo.fingerprint.com/playground — should not show "nodriver" or "bot" detection.https://antcpt.com/eng/information/demo-form/recaptcha-3-test-score.html — score should be ≥ 0.7.https://deviceandbrowserinfo.com — isBot should be false with 0 true flags.navigator.webdriver is false with page.evaluate("navigator.webdriver").humanize-automation — Human-like mouse/keyboard/scroll for behavioral bypass.tls-fingerprint-impersonation — TLS/JA4 fingerprint spoofing at the HTTP client level.http2-header-impersonation — Browser-specific HTTP/2 pseudo-header ordering and SETTINGS frames.© uphiago, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
Just SKILL.md in recon/stealth-browser-launch of uphiago/recon-skills.
Open the folder on GitHubat commit 1260244
Stealth Browser Launch next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Stealth Browser Launch this skilluphiago/recon-skills | 1.3k | — | ~2.2k | Automated safety check: Pass | MIT | |
| ONNX Runtime Shape Inference Safety Auditmicrosoft/onnxruntime | 22k | — | ~3.3k | Automated safety check: Pass | MIT | |
| Code Audit3stoneBrother/code-audit | 893 | 1 repos | ~2.7k | Automated safety check: Pass | None | |
| CodeQL Security Scantrailofbits/skills | 7.4k | — | ~4.6k | Automated safety check: Notes | CC-BY-SA-4.0 | |
| Doc CommentsMichaelGrafnetter/DSInternals | 2k | — | ~1.2k | Automated safety check: Pass | MIT | |
| Harness Design Fuzzingprovos/ironcurtain | 613 | — | ~5.7k | Automated safety check: Pass | Apache-2.0 |
microsoft/onnxruntime
Finds and fixes out-of-range output writes in ONNX Runtime operator shape-inference functions where a getNumOutputs guard admits too few outputs.
3stoneBrother/code-audit
Professional code security audit skill covering 55+ vulnerability types.
trailofbits/skills
Scans a codebase for vulnerabilities with CodeQL's data flow and taint tracking in run-all or important-only modes, including data extensions for project-specific sources and sinks.
MichaelGrafnetter/DSInternals
Ensure that C and C++/CLI types are documented with XML comments and follow best practices for documentation.
provos/ironcurtain
Reference vocabulary for designing instrumented harnesses that drive vulnerability discovery — design classes (trigger-driven vs coverage-driven), tiered scope (T1 isolated function / T2…
hardw00t/ai-security-arsenal
Automated Dynamic Application Security Testing (DAST) using Playwright MCP plus standard OS pentest tooling.
uphiago/recon-skills
Flags API endpoints whose data or actions look like they should need a login but currently don't, as part of authorized security testing.
uphiago/recon-skills
Mine errorlog for creds, paths, SQL when leak hunt finds. An agent skill from uphiago/recon-skills.
uphiago/recon-skills
Analyze JS bundles and source maps for hardcoded secrets, API keys, JWTs, and internal endpoints
uphiago/recon-skills
A skill your agent uses when starting or restructuring an authorized external web and API assessment.
uphiago/recon-skills
Sensitive file scanning, path traversal bypass, vHost enum, .env extract, log mining, Varnish detect
uphiago/recon-skills
A skill your agent uses when protected HTTP routes return 401 or 403.
Works with
Categories
Launch stealth Chromium with C++ fingerprint patches for anti-bot bypass. Stealth Browser Launch is an agent skill from uphiago/recon-skills. Launch stealth Chromium with C++ fingerprint patches for anti-bot bypass.
Stealth Browser Launch fits situations like: security work in your project.
Run `npx skills add uphiago/recon-skills --skill stealth-browser-launch -a claude-code`. Or copy the skill folder (recon/stealth-browser-launch in uphiago/recon-skills) into .claude/skills/stealth-browser-launch in your project. Claude Code loads it when a task matches its description.
Run `npx skills add uphiago/recon-skills --skill stealth-browser-launch -a codex`. Or copy the skill folder (recon/stealth-browser-launch in uphiago/recon-skills) into .agents/skills/stealth-browser-launch in your project. Codex loads it when a task matches its description.
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add uphiago/recon-skills --skill stealth-browser-launch -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/stealth-browser-launch, .gemini/skills/stealth-browser-launch, .github/skills/stealth-browser-launch and .opencode/skills/stealth-browser-launch in your project.
Going by SKILL.md and its folder, Stealth Browser Launch needs the command-line tools its instructions call (pip, python3, docker, apt and playwright). Our summary lists: Python 3; Docker. Compatibility (from SKILL.md): Requires curl, httpx, nuclei, python3.
SKILL.md names 4 domains. In commands or code: browserscan.net, demo.fingerprint.com, antcpt.com and deviceandbrowserinfo.com; the agent is likely to contact these when it follows the instructions. This is read from the text; nothing was executed.
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.
Stealth Browser Launch is published under the MIT licence (declared in SKILL.md). It allows redistribution, so the full SKILL.md is shown on this page.
About 2.2k tokens (SKILL.md is roughly 8.6k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.
Skills that share tags, products or a category with Stealth Browser Launch: ONNX Runtime Shape Inference Safety Audit (microsoft/onnxruntime, 22k stars), Code Audit (3stoneBrother/code-audit, 893 stars), CodeQL Security Scan (trailofbits/skills, 7.4k stars) and Doc Comments (MichaelGrafnetter/DSInternals, 2k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
uphiago (a GitHub user) maintains it in uphiago/recon-skills, which has 1,294 GitHub stars. The repository holds 23 skills in this directory. The repository was last updated on September 1, 2026.
Source: uphiago/recon-skills on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.