Back to skill

Security audit

Agent Browser Backup

Security checks for vulnerabilities and agentic risk

Overview

The skill is a straightforward browser-automation wrapper, but users should treat saved sessions, cookies, screenshots, recordings, and unpinned installs as sensitive.

Install only from a trusted reviewed version if possible. Do not save auth state, screenshots, PDFs, traces, or recordings from sensitive sites unless needed, and delete or protect those files afterward. Avoid putting real passwords or tokens directly in commands that may be logged.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Findings (1)

T08 · Insecure Dependencies

Warning
Location
SKILL.md:16
Finding

Mutable Third-Party Dependencies Are Installed and Executed Without Version or Commit Pinning

Content
View full analysis

Vulnerability Details

File Location: SKILL.md:16-32; CONTRIBUTING.md:23-26
Vulnerability Type: Unpinned third-party package and source dependencies
Risk Level: Medium

Vulnerable Code

SKILL.md:16-32:

bash
### npm recommended

```bash
npm install -g agent-browser
agent-browser install
agent-browser install --with-deps

From Source

bash
git clone https://github.com/vercel-labs/agent-browser
cd agent-browser
pnpm install
pnpm build
agent-browser install
text

`CONTRIBUTING.md:23-26`:

```bash
1. Install the latest version
   ```bash
   npm install -g agent-browser@latest
   ```

Technical Analysis

The installation instructions retrieve and execute mutable third-party content. The npm commands install either the package's current default release or explicitly use the mutable latest distribution tag. The source installation procedure clones the repository's current default branch without checking out a reviewed commit or verifying a signed release.

Package installation and source builds can execute npm/pnpm lifecycle and build scripts. Consequently, the code executed by these instructions may differ from the code that was available when this Skill was audited. A malicious or compromised future npm release, repository commit, maintainer account, or transitive dependency could therefore introduce arbitrary executable behavior.

The referenced package and repository names are consistent with the documented browser automation project. The audited files do not demonstrate typosquatting, dependency confusion, or an existing malicious payload. The risk arises from the absence of immutable version pinning and integrity verification.

Attack Path

  1. An attacker compromises the npm package, its publisher account, the upstream repository, or a relevant transitive dependency.
  2. The attacker publishes a malicious release under the default or `latest ...[truncated 1160 chars]
Remediation
View remediation

Remediation Suggestions

  1. Pin agent-browser to an exact, reviewed npm version rather than relying on the default release or @latest:
    bash
    npm install -g agent-browser@<reviewed-exact-version>
    
  2. For source installations, check out an immutable reviewed commit SHA:
    bash
    git clone https://github.com/vercel-labs/agent-browser
    cd agent-browser
    git checkout --detach <reviewed-commit-sha>
    
  3. Verify signed tags, release provenance, and published package integrity before installation.
  4. Use a committed lockfile with frozen-lockfile installation for source builds so transitive dependency resolution cannot change silently.
  5. Review package lifecycle and build scripts before execution. Where operationally possible, initially install dependencies with lifecycle scripts disabled and explicitly run only reviewed setup steps.
  6. Execute installation and browser automation in a least-privileged container or sandbox without unnecessary secrets, host filesystem access, or elevated permissions.
  7. Replace the @latest troubleshooting recommendation in CONTRIBUTING.md with a known-compatible exact version and document a controlled process for reviewing upgrades.
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Memory PoisoningPersistent Context Injection, Context Window Stuffing, Memory Manipulation
  • Rogue AgentSelf-Modification, Session Persistence
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (6)

Self-Modification

High
Category
Rogue Agent
Confidence
85% confidence
Finding

Skill modifies its own code, configuration, or behavior at runtime. Self-modification enables an agent to escalate privileges, disable safety constraints, or install persistent backdoors.

Content

Scanner excerpt · CONTRIBUTING.md (reported line 60)May include surrounding context.

md
## Adding New Commands to the Skill

Update SKILL.md when the upstream CLI adds new commands.
- Keep the Installation section
- Add new commands in the correct category
- Include usage examples

Context Window Stuffing

Medium
Category
Memory Poisoning
Confidence
85% confidence
Finding

Skill attempts to fill the context window with filler content, displacing legitimate instructions and safety constraints. This can degrade agent performance or bypass safety boundaries.

Content

Scanner excerpt · SKILL.md (reported line 41)May include surrounding context.

agent-browser open # Navigate to page agent-browser snapshot -i # Get interactive elements with refs agent-browser click @e1 # Click element by ref agent-browser fill @e2 "text" # Fill input by ref agent-browser close # Close browser

text

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The skill explicitly documents commands for setting HTTP credentials and manipulating cookies/localStorage without any warning about secret handling, token exposure, or safe storage. In an agent context, this can normalize passing credentials on the command line and persisting sensitive session material where it may be logged, echoed, or reused unintentionally.

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
95% confidence
Finding

The skill documents saving and loading browser session state to a file, which can preserve cookies, tokens, and other authenticated state. Without warnings or safeguards, agents may create reusable authentication artifacts that enable account takeover or unauthorized session reuse if the file is exposed.

Content

Scanner excerpt · SKILL.md (reported line 236)May include surrounding context.

bash
agent-browser state save auth.json    # Save session state
agent-browser state load auth.json    # Load saved state

Example: Form submission

Session Persistence

Medium
Category
Rogue Agent
Confidence
96% confidence
Finding

The authentication example encourages saving a logged-in session to auth.json and reloading it later, directly demonstrating persistence of authenticated state. In the context of an agent skill, this materially increases risk because it operationalizes credential/session reuse without discussing theft, replay, or cross-task leakage.

Content

Scanner excerpt · SKILL.md (reported line 265)May include surrounding context.

agent-browser wait --url "/dashboard" agent-browser state save auth.json

Later sessions: load saved state

agent-browser state load auth.json agent-browser open https://app.example.com/dashboard

text

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
84% confidence
Finding

The documentation advertises screenshots, PDFs, and video recording that write artifacts to disk, but gives no warning that these files may capture sensitive page contents, credentials, or personal data. In an automated agent workflow, silent local persistence increases the risk of data leakage through leftover files, workspace sharing, or later exfiltration by other tools.

Content

No source excerpt is available for this finding.