Back to skill

Security audit

OpenClaw Vertex Setup

Security checks for vulnerabilities and agentic risk

Overview

This looks like a real OpenClaw Vertex setup guide, but it asks users to broaden credential exposure and refresh a persistent Gateway service without enough safety scoping.

Review this skill before installing. Prefer a pinned OpenClaw npm version, verify the active `openclaw` binary before reinstalling the Gateway, back up existing service/config files, keep `~/.openclaw/.env` and Google ADC files readable only by your user, and avoid enabling broad shell environment import unless you understand which secrets your shell exports.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (2)

T08 · Insecure Dependencies

Warning
Location
SKILL.md:103
Finding

Unpinned Global npm Package Installation Creates a Supply-Chain Execution Risk

Content
View full analysis
Remediation
View remediation
``` 2. Document the expected package version and published integrity digest. 3. Verify npm provenance, package ownership, and integrity before installation. 4. Review release notes and package lifecycle scripts before upgrading. 5. Avoid automatically reinstalling the Gateway immediately after an unreviewed upgrade. 6. Validate the installed executable and package version before registering the service: ```bash which openclaw openclaw --version npm list -g --depth=0 openclaw ``` 7. Where feasible, test the pinned release in an isolated environment before deploying it as a persistent Gateway service. ]]>

T09 · Insecure Skill Coding Practices

Warning
Location
SKILL.md:117
Finding

Broad Shell Environment Import Exposes Unrelated Secrets to the Gateway Process

Content
View full analysis
Remediation
View remediation
GOOGLE_CLOUD_LOCATION=global GOOGLE_APPLICATION_CREDENTIALS=/Users//.config/gcloud/application_default_credentials.json ``` 3. Use an explicit environment-variable allowlist if OpenClaw supports one. 4. Ensure `~/.openclaw/.env` and the ADC credential file are readable only by the owning user, for example: ```bash chmod 600 ~/.openclaw/.env chmod 600 ~/.config/gcloud/application_default_credentials.json ``` 5. Avoid exporting unrelated long-lived secrets from global shell startup files. 6. Prefer short-lived credentials and narrowly scoped service identities where supported. 7. Document broad shell import only as an opt-in compatibility fallback, accompanied by a warning that it may expose unrelated credentials to the Gateway process. ]]>
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Rogue AgentSelf-Modification, Session Persistence
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (24)

Credential Access

High
Category
Privilege Escalation
Confidence
80% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · README.md (reported line 55)May include surrounding context.

dotenv
GOOGLE_CLOUD_PROJECT=<project-id>
GOOGLE_CLOUD_LOCATION=global
GOOGLE_APPLICATION_CREDENTIALS=/Users/<user>/.config/gcloud/application_default_credentials.json

3. 在 ~/.openclaw/openclaw.json 里设置默认模型

Credential Access

High
Category
Privilege Escalation
Confidence
80% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · README.zh-CN.md (reported line 55)May include surrounding context.

dotenv
GOOGLE_CLOUD_PROJECT=<project-id>
GOOGLE_CLOUD_LOCATION=global
GOOGLE_APPLICATION_CREDENTIALS=/Users/<user>/.config/gcloud/application_default_credentials.json

3. 在 ~/.openclaw/openclaw.json 里设置默认模型

Credential Access

High
Category
Privilege Escalation
Confidence
80% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · SKILL.md (reported line 62)May include surrounding context.

dotenv
GOOGLE_CLOUD_PROJECT=<project-id>
GOOGLE_CLOUD_LOCATION=global
GOOGLE_APPLICATION_CREDENTIALS=/Users/<user>/.config/gcloud/application_default_credentials.json

3. 在 ~/.openclaw/openclaw.json 里设置默认模型

Credential Access

High
Category
Privilege Escalation
Confidence
80% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · SKILL.md (reported line 218)May include surrounding context.

dotenv
GOOGLE_CLOUD_PROJECT=<project-id>
GOOGLE_CLOUD_LOCATION=global
GOOGLE_APPLICATION_CREDENTIALS=/Users/<user>/.config/gcloud/application_default_credentials.json

3. 在 ~/.openclaw/openclaw.json 里设置默认模型

Credential Access

High
Category
Privilege Escalation
Confidence
80% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · README.md (reported line 55)May include surrounding context.

md
- Gateway service wiring: `openclaw gateway status --deep`
- launchd service file: `~/Library/LaunchAgents/ai.openclaw.gateway.plist`
- Shell env: `~/.zshrc`
- ADC file: `~/.config/gcloud/application_default_credentials.json`

Confirm the default OpenClaw model points at Vertex:

Credential Access

High
Category
Privilege Escalation
Confidence
80% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · README.zh-CN.md (reported line 55)May include surrounding context.

md
- Gateway service wiring: `openclaw gateway status --deep`
- launchd service file: `~/Library/LaunchAgents/ai.openclaw.gateway.plist`
- Shell env: `~/.zshrc`
- ADC file: `~/.config/gcloud/application_default_credentials.json`

Confirm the default OpenClaw model points at Vertex:

Credential Access

High
Category
Privilege Escalation
Confidence
80% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · SKILL.md (reported line 33)May include surrounding context.

md
- Gateway service wiring: `openclaw gateway status --deep`
- launchd service file: `~/Library/LaunchAgents/ai.openclaw.gateway.plist`
- Shell env: `~/.zshrc`
- ADC file: `~/.config/gcloud/application_default_credentials.json`

Confirm the default OpenClaw model points at Vertex:

Credential Access

High
Category
Privilege Escalation
Confidence
80% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · SKILL.md (reported line 62)May include surrounding context.

md
- Gateway service wiring: `openclaw gateway status --deep`
- launchd service file: `~/Library/LaunchAgents/ai.openclaw.gateway.plist`
- Shell env: `~/.zshrc`
- ADC file: `~/.config/gcloud/application_default_credentials.json`

Confirm the default OpenClaw model points at Vertex:

Credential Access

High
Category
Privilege Escalation
Confidence
80% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · SKILL.md (reported line 218)May include surrounding context.

md
- Gateway service wiring: `openclaw gateway status --deep`
- launchd service file: `~/Library/LaunchAgents/ai.openclaw.gateway.plist`
- Shell env: `~/.zshrc`
- ADC file: `~/.config/gcloud/application_default_credentials.json`

Confirm the default OpenClaw model points at Vertex:

Credential Access

High
Category
Privilege Escalation
Confidence
70% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · README.md (reported line 55)May include surrounding context.

md
- Gateway service wiring: `openclaw gateway status --deep`
- launchd service file: `~/Library/LaunchAgents/ai.openclaw.gateway.plist`
- Shell env: `~/.zshrc`
- ADC file: `~/.config/gcloud/application_default_credentials.json`

Confirm the default OpenClaw model points at Vertex:

Credential Access

High
Category
Privilege Escalation
Confidence
70% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · README.zh-CN.md (reported line 55)May include surrounding context.

md
- Gateway service wiring: `openclaw gateway status --deep`
- launchd service file: `~/Library/LaunchAgents/ai.openclaw.gateway.plist`
- Shell env: `~/.zshrc`
- ADC file: `~/.config/gcloud/application_default_credentials.json`

Confirm the default OpenClaw model points at Vertex:

Credential Access

High
Category
Privilege Escalation
Confidence
70% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · SKILL.md (reported line 33)May include surrounding context.

md
- Gateway service wiring: `openclaw gateway status --deep`
- launchd service file: `~/Library/LaunchAgents/ai.openclaw.gateway.plist`
- Shell env: `~/.zshrc`
- ADC file: `~/.config/gcloud/application_default_credentials.json`

Confirm the default OpenClaw model points at Vertex:

Credential Access

High
Category
Privilege Escalation
Confidence
70% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · SKILL.md (reported line 62)May include surrounding context.

md
- Gateway service wiring: `openclaw gateway status --deep`
- launchd service file: `~/Library/LaunchAgents/ai.openclaw.gateway.plist`
- Shell env: `~/.zshrc`
- ADC file: `~/.config/gcloud/application_default_credentials.json`

Confirm the default OpenClaw model points at Vertex:

Credential Access

High
Category
Privilege Escalation
Confidence
70% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · SKILL.md (reported line 218)May include surrounding context.

md
- Gateway service wiring: `openclaw gateway status --deep`
- launchd service file: `~/Library/LaunchAgents/ai.openclaw.gateway.plist`
- Shell env: `~/.zshrc`
- ADC file: `~/.config/gcloud/application_default_credentials.json`

Confirm the default OpenClaw model points at Vertex:

Credential Access

High
Category
Privilege Escalation
Confidence
80% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · SKILL.md (reported line 33)May include surrounding context.

md
- Gateway service wiring: `openclaw gateway status --deep`
- launchd service file: `~/Library/LaunchAgents/ai.openclaw.gateway.plist`
- Shell env: `~/.zshrc`
- ADC file: `~/.config/gcloud/application_default_credentials.json`

Confirm the default OpenClaw model points at Vertex:

Credential Access

High
Category
Privilege Escalation
Confidence
80% confidence
Finding

The issue is not the filename itself, but that the skill advises persisting a credential-related file reference in a shared OpenClaw environment file without secure-handling guidance. That can create avoidable exposure of cloud-auth configuration across service contexts and local users.

Content

Scanner excerpt · SKILL.md (reported line 86)May include surrounding context.

dotenv
GOOGLE_CLOUD_PROJECT=<project-id>
GOOGLE_CLOUD_LOCATION=global
GOOGLE_APPLICATION_CREDENTIALS=/Users/<user>/.config/gcloud/application_default_credentials.json

This is the most reliable fix for:

Credential Access

High
Category
Privilege Escalation
Confidence
80% confidence
Finding

The issue is not the filename itself, but that the skill advises persisting a credential-related file reference in a shared OpenClaw environment file without secure-handling guidance. That can create avoidable exposure of cloud-auth configuration across service contexts and local users.

Content

Scanner excerpt · SKILL.md (reported line 86)May include surrounding context.

dotenv
GOOGLE_CLOUD_PROJECT=<project-id>
GOOGLE_CLOUD_LOCATION=global
GOOGLE_APPLICATION_CREDENTIALS=/Users/<user>/.config/gcloud/application_default_credentials.json

This is the most reliable fix for:

Credential Access

High
Category
Privilege Escalation
Confidence
76% confidence
Finding

The issue is not the filename itself, but that the skill advises persisting a credential-related file reference in a shared OpenClaw environment file without secure-handling guidance. That can create avoidable exposure of cloud-auth configuration across service contexts and local users.

Content

Scanner excerpt · SKILL.md (reported line 86)May include surrounding context.

dotenv
GOOGLE_CLOUD_PROJECT=<project-id>
GOOGLE_CLOUD_LOCATION=global
GOOGLE_APPLICATION_CREDENTIALS=/Users/<user>/.config/gcloud/application_default_credentials.json

This is the most reliable fix for:

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
83% confidence
Finding

The README instructs users to place a Google application credentials path in a shared OpenClaw environment file under ~/.openclaw without any warning about access controls or secret-handling implications. Even though it stores a path rather than the credential contents, centralizing auth configuration in a broadly reused file can unintentionally expose or over-share sensitive authentication material to other tools, agents, logs, or users on the same system.

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
75% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · SKILL.md (reported line 31)May include surrounding context.

md
- Agent auth store: `~/.openclaw/agents/main/agent/auth-profiles.json`
- Active CLI path: `which -a openclaw`
- Gateway service wiring: `openclaw gateway status --deep`
- launchd service file: `~/Library/LaunchAgents/ai.openclaw.gateway.plist`
- Shell env: `~/.zshrc`
- ADC file: `~/.config/gcloud/application_default_credentials.json`

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The skill instructs the user to persist cloud auth-related environment variables, including the path to application default credentials, into a shared ~/.openclaw/.env file without any warning about file permissions, secret exposure through logs, or who can read that file. While it does not exfiltrate credentials, it normalizes storing sensitive auth configuration in a broadly reused location tied to a service process, which can increase accidental disclosure risk on multi-user systems or through debugging output.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
85% confidence
Finding

The README instructs users to place a path to their ADC credential file in a shared OpenClaw config file without any warning about file permissions, exposure risk, or whether the config may be read by other tools, plugins, or logs. While it does not disclose the credential contents itself, normalizing credential references in broadly used config increases the chance of accidental leakage or unsafe sharing of a sensitive auth setup.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
72% confidence
Finding

This file presents its instructions in a single enforced language variant, which can be a locale/language policy concern when users are not given an in-file choice. While it links to an English version, the content here itself does not offer language selection beyond the heading link.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
86% confidence
Finding

The README recommends forceful Gateway reinstall and restart commands that modify local launch/service configuration, but it does not warn that these actions can overwrite an existing setup or disrupt a running environment. In a system administration skill, this creates operational risk and can cause accidental service breakage or confusing state changes during troubleshooting.

Content

No source excerpt is available for this finding.