agent-experience

81 / 100 · Low risk

Fine to try. Some of what it can do is expected: a security skill usually runs scanners and checks websites.

By browserbase/skills (GitHub) · security · 3,756 stars · scanned at commit 6811ca3

Audit the developer experience of a product, SDK, docs site, or SKILL.md by dropping multiple Claude subagents at it with only a tiny task prompt and real tools (WebFetch, Bash, Write). Agents must discover the docs themselves, install deps, ask for credentials if needed, and attempt real execution. The skill captures each agent's trace — tool calls, retries, wall time, errors — and scores on Setup Friction, Speed, Efficiency, Error Recovery, and Doc Quality, then emits an HTML report with an A…

Works with any agent that supports SKILL.md skills.

What this skill can do

Runs commands on your computer

This is disclosed by the author in the skill's SKILL.md or README. Expected for a security skill.

Found in 3 places.

skills/agent-experience/SKILL.md, line 119printenv | grep -iE '^[A-Z][A-Z0-9_]*_(API_KEY|TOKEN|SECRET|KEY)=' | cut -d= -f1
skills/agent-experience/SKILL.md, line 127grep -hE '^[[:space:]]*export[[:space:]]+[A-Z][A-Z0-9_]*_(API_KEY|TOKEN|SECRET|KEY)=' \
skills/agent-experience/SKILL.md, line 423rm -rf ./dx-audit-tmp/

Reads your browser cookies or saved passwords

This is disclosed by the author in the skill's SKILL.md or README.

Found in 2 places.

skills/agent-experience/SKILL.md, line 140- macOS Keychain: `security find-generic-password -l "<expected-name>" -w` — try once with the most likely name (e.g. `BROWSERBASE_API_KEY`); silent failure mea
skills/agent-experience/SKILL.md, line 158- Tier 3: `op read "op://<vault>/<item>/<field>"` or `security find-generic-password -l <NAME> -w`.

Sends data to these websites: api.browserbase.com, docs.example.com

This is disclosed by the author in the skill's SKILL.md or README. Expected for a security skill.

Found in 9 places.

skills/agent-experience/SKILL.md, line 205- Use your real tools (`WebFetch`, `Bash` if allowed, `Write`)
skills/agent-experience/SKILL.md, line 217**Verification of agent claims using orchestrator credentials.** Before scoring, if Step 2.5 retained product-specific credentials, the orchestrator may use the
skills/agent-experience/SKILL.md, line 237Classify each `WebFetch` URL into one of four provenance categories and render with the matching label + color:

Show 6 more

skills/agent-experience/SKILL.md, line 348[+1.2s] [TOOL_USE #1] WebFetch
skills/agent-experience/SKILL.md, line 372- Input/Output: visible inline as `.trace-io` blocks with colored left-border (orange for input, green for output, red for errors). `<pre>` block shows the **fu
skills/agent-experience/references/subagent-brief.md, line 18- Use your real tools: WebFetch to read docs, Read/Write for files, {{EXEC_MODE_TOOLS}}.
skills/agent-experience/references/subagent-brief.md, line 53{"tool": "WebFetch", "count": 4, "purpose": "fetched docs pages"},
skills/agent-experience/references/subagent-brief.md, line 73{"t_ms": 1200, "type": "tool_use", "n": 1, "tool": "WebFetch", "input": {"url": "https://docs.example.com", "prompt": "What is this?"}},
skills/agent-experience/references/subagent-brief.md, line 150- `{{EXEC_MODE_NOTE}}` → `Do not run any Bash commands. Draft the code you would run, list the commands you would execute, and score the docs based on what you

Install

Paste this into your terminal. It downloads exactly the version we scanned, not whatever is newest.

if [ -e ~/.agents/skills/'agent-experience' ] || [ -L ~/.agents/skills/'agent-experience' ]; then echo 'A skill folder named agent-experience already exists — not replacing it' >&2; false; else d=$(mktemp -d) && git clone --quiet --filter=blob:none --no-checkout 'https://github.com/browserbase/skills' "$d" && git -C "$d" checkout --quiet '6811ca31163332d9d60309cff48e77f09de37a17' -- 'skills/agent-experience' && mkdir -p ~/.agents/skills && cp -R "$d"/'skills/agent-experience' ~/.agents/skills/'agent-experience' && rm -rf "$d"; fi

Codex reads ~/.agents/skills directly. For Claude Code, Cursor, Copilot, Gemini CLI, Antigravity or OpenCode, also link it. For Claude Code: ln -s ~/.agents/skills/'agent-experience' ~/.claude/skills/'agent-experience' (other agents: ~/.cursor/skills, ~/.copilot/skills, ~/.gemini/skills, ~/.gemini/config/skills, ~/.config/opencode/skills). Or install through the finder skill, which does both.

Badge for your README

The badge for browserbase/skills shows the score of its lowest-scoring skill.

SkillTrust badge for browserbase/skills

[![SkillTrust](https://zalavijaysinh.github.io/skilltrust/badges/browserbase/skills.svg)](https://zalavijaysinh.github.io/skilltrust/repos/browserbase/skills/)

Score history

Date	Version	Score
2026-11-08	`6811ca3`	81
2026-11-06	`6811ca3`	71