Skills DirectorySkills Directory
SkillsLearnSecurityCategoriesDocsBlogPro
Sign InSubmit Skill
Skills Directory

Security-tested agent skills for Claude, coding agents, and AI workflows.

Directory

  • Browse Skills
  • All Skills A–Z
  • Claude Skills
  • Claude Code Skills
  • Agent Skills
  • Categories
  • Authors
  • Submit a Skill

Learn

  • Learn Hub
  • Install Claude Skills
  • Write SKILL.md
  • Skills vs MCP
  • Directories Compared

Security

  • Security
  • Methodology
  • Secure Claude Skills
  • Security Badges
  • Chrome Extension
  • Skill Manager

Company

  • About
  • Community
  • Blog
  • API Docs
  • Advertise

2026 Skills Directory. All rights reserved.

ProTermsPrivacyRefunds
Back to skills

Jev Browser Choice

ASecurity

Use Jev selectively to choose among safe, observed controls in an authorized host browser tab; works with the current native browser interface without a separate browser runtime.

2 stars
0 votes
0 copies
0 views
Added 9/30/2026
ai-agentsgo

Works with

cli

Security Analysis

A100/100

Pro scans all 2 files and shows the line behind each finding

Scanned 10/1/2026

$npx -y skills add qualixar/jev-decision-layer --skill jev-browser-choice --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Jev Browser Choice?

Add the live security badge to your README — it updates automatically with every re-scan.

Security grade badge for Jev Browser Choice
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/qualixar-jev-browser-choice/badge)](https://www.skillsdirectory.com/skills/qualixar-jev-browser-choice)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
Files
SKILL.md
---
name: jev-browser-choice
description: Use Jev selectively to choose among safe, observed controls in an authorized host browser tab; works with the current native browser interface without a separate browser runtime.
---

# Selective browser choice

Use the host's available native browser or Computer Use interface to observe the live tab. The user and host authorize browser access; this skill grants no new permission. Never pass cookies, credentials, hidden fields, full-page dumps, screenshots, or account details to Jev.

**No Jev call** when one visible safe control clearly serves the goal, or when a deterministic rule already decides it. For example, a cart-flow test calls for “Add to Cart,” not “Buy Now.” Calling a model for that choice adds overhead. Use Jev only when at least two plausible, safe observed controls need semantic judgment and a wrong turn could cost more work than one bounded request.

For a single ambiguous choice, call `jev_route` with `kind: "tool"`, a concise goal, and 2–12 candidates copied from the current visible controls, each exactly `{"id", "description"}` with a short id that starts with a letter and has no dots or colons. Do not add an `unknown` candidate yourself: the tool always offers it, and an id of `unknown` is refused. Pass the `workspace_path` and `data_classification` the workspace's grant accepts. For a multi-step page where both the operation and target are genuinely uncertain, one `jev_typed_decide` request may ask parallel Choice questions over the same compact observed state: operation (`CLICK`, `SCROLL`, `WAIT`, `STOP`) and a compatible target ID. Ignore an unused target head. Do not make sequential confirmation calls for the same unchanged page.

Keep an in-turn table from each candidate ID to its *observed* native browser control. A Jev answer is advice, never a selector, coordinate, script, permission, or instruction to execute. Reject unknown, blocked, unavailable, unsafe, or stale choices. Before acting, re-observe enough of the page to confirm that the selected control still exists with the same label and purpose; after acting, re-observe the result. If the page changed, rebuild the table or hand control back to the host—never click an old element index.

Do not offer purchase, checkout, payment, credential, account-permission, or irreversible submission controls to Jev as executable candidates. The host's normal confirmation rules still apply to any consequential action, including an explicitly requested cart mutation. Text entry uses only user-provided or independently verified text through the native browser tool; Jev does not generate form values. Stop when the stated browser goal is visibly verified, and preserve the original browser state when the task calls for a temporary test.

The older `bridge.mjs` remains an optional, navigation-link-only compatibility artifact for environments that support it; it cannot auto-click cart, form, account, checkout, or payment controls. It is not required for this skill or evidence of current Codex Desktop browser interception. A smaller decision payload may reduce model context, but token, cost, and time savings require matched end-to-end measurements.

Attribution

qualixarqualixar
View sourceSee grades on GitHubMore from qualixar →
SSkills DirectorySkills Directory

Ship a skill? Prove it's safe.

Free 120-pattern security scan, letter grade, and an embeddable README badge.

Submit a skill

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments (0)

No comments yet. Be the first to comment!

SSkills DirectorySkills Directory

Ship a skill? Prove it's safe.

Free 120-pattern security scan, letter grade, and an embeddable README badge.

Submit a skill

Related Skills

Caveman

Terse caveman voice: answer first, fluff gone, every technical fact kept. Use for /caveman, "caveman mode", "talk like caveman", "be brief", "less tokens". Stays on until "stop caveman" or "normal mode".

1100021 votes

Hyperplan

Adversarial multi-agent planning skill. Self-orchestrates 5 hostile category members (unspecified-low, unspecified-high, deep, ultrabrain, artistry) via team-mode for ruthless cross-critique debate, distills only the defensible insights, then MANDATORILY hands the distilled insight bundle to the `plan` agent for executable plan formalization. Use when planning needs maximum rigor and surfacing of weak assumptions, blind spots, and over-engineering. Triggers: 'hyperplan', 'hpp', '/hyperplan', ...

698431 votes

Writing Skills

Create and manage Claude Code skills in HASH repository following Anthropic best practices. Use when creating new skills, modifying skill-rules.json, understanding trigger patterns, working with hooks, debugging skill activation, or implementing progressive disclosure. Covers skill structure, YAML frontmatter, trigger types (keywords, intent patterns), UserPromptSubmit hook, and the 500-line rule. Includes validation and debugging with SKILL_DEBUG. Examples include rust-error-stack, cargo-dep...

3931 votes

Mcp Code Execution

Routes multi-tool workflows through MCP servers for large datasets and pipelines. Use when Bash tool overhead is limiting throughput on data-heavy tasks.

3421 votes

catchup

Recovers the conversation and failed tool calls of a previous Codex, Amp, Claude Code, Antigravity, Cline, Copilot CLI, Cursor, DeepSeek Harness, Grok Build, Kimi, OpenCode, Pi Agent, or ZCode session. Use when the user says "catch up", "what did the last session do", "get me up to speed", "I switched agents", asks to recover/summarize a previous session before continuing, or asks to diagnose or report a catchup failure. Do NOT use for the current conversation, git history, or any non-agent log.

741 votes
View all in ai-agents →