Back to skill

Security audit

Coding Plan Assistant

Security checks for vulnerabilities and agentic risk

Overview

The skill is mostly coherent, but it handles API keys with weak local safeguards and overstates how safely they are stored.

Review this before installing if you plan to store real API keys. Prefer a dedicated secret manager or OS keychain, avoid passing keys on the command line, manually protect or ignore .openclaw/.env, and verify any npm install commands from the providers' official documentation.

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 (3)

T09 · Insecure Skill Coding Practices

Error
Location
index.js:72
Finding

API Keys Are Stored in a Plaintext File Without Restrictive Permissions

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Error
Location
index.js:267
Finding

API Keys Can Be Passed Through Process Command-Line Arguments

Content
View full analysis
Remediation
View remediation

T08 · Insecure Dependencies

Warning
Location
references/codex-openai.md:25
Finding

Documentation Recommends Unpinned Third-Party npm Installations

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Findings (44)

Tp4

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The skill combines API key management, status inspection, and rotation workflows while presenting itself as a planning assistant. For a secret-handling tool, misleading framing is especially risky because users may not apply the scrutiny they would to a credential manager, making accidental disclosure or unauthorized changes more likely.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding

The skill combines API key management, status inspection, and rotation workflows while presenting itself as a planning assistant. For a secret-handling tool, misleading framing is especially risky because users may not apply the scrutiny they would to a credential manager, making accidental disclosure or unauthorized changes more likely.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding

The skill combines API key management, status inspection, and rotation workflows while presenting itself as a planning assistant. For a secret-handling tool, misleading framing is especially risky because users may not apply the scrutiny they would to a credential manager, making accidental disclosure or unauthorized changes more likely.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The skill combines API key management, status inspection, and rotation workflows while presenting itself as a planning assistant. For a secret-handling tool, misleading framing is especially risky because users may not apply the scrutiny they would to a credential manager, making accidental disclosure or unauthorized changes more likely.

Content

No source excerpt is available for this finding.

Credential Access

High
Category
Privilege Escalation
Confidence
91% confidence
Finding

The skill is designed to access a local .env file containing multiple API keys, which is direct interaction with highly sensitive credentials. Even if the docs mention masking and .gitignore, centralizing read/write/rotation of secrets inside a broadly described skill creates substantial risk of exfiltration, accidental disclosure, or unauthorized overwrite if triggered improperly.

Content

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

鍑嵁鏍煎紡

env
# .openclaw/.env
GITHUB_COPILOT_TOKEN=ghp_xxxx...xxxx
CLAUDE_API_KEY=sk-ant-xxxx...xxxx
OPENAI_API_KEY=sk-xxxx...xxxx

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 106)May include surrounding context.

md
### scripts/check-status.js

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 109)May include surrounding context.

md
### scripts/check-status.js

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 112)May include surrounding context.

md
### scripts/rotate-key.js

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 115)May include surrounding context.

md
### scripts/rotate-key.js

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 157)May include surrounding context.

md
1. 鍦?`config.json` 涓坊鍔犲钩鍙伴厤缃?2. 鍦?`references/` 涓垱寤哄畼鏂规枃妗i摼鎺?3. 鏇存柊 `SKILL.md` 鐨勫钩鍙版竻鍗?

Credential Access

High
Category
Privilege Escalation
Confidence
60% confidence
Finding

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

Content

Scanner excerpt · config.json (reported line 134)May include surrounding context.

json
}
  },
  "security": {
    "envPath": ".openclaw/.env",
    "maskPattern": "...***",
    "rotationReminder": "90 天"
  },

Credential Access

High
Category
Privilege Escalation
Confidence
85% confidence
Finding

The skill is explicitly designed to access a shared .openclaw/.env path under the invoking working directory, creating a capability to read and later expose whether credentials exist across platforms. In agent ecosystems, direct filesystem access to secret files is sensitive because it can be repurposed to enumerate or misuse credentials outside a narrowly scoped secret-management boundary.

Content

Scanner excerpt · index.js (reported line 16)May include surrounding context.

js
// 技能目录
const SKILL_DIR = __dirname;
const CONFIG_PATH = path.join(SKILL_DIR, 'config.json');
const ENV_PATH = path.join(process.env.INIT_CWD || process.cwd(), '.openclaw', '.env');

// 加载配置
let config = {};

Credential Access

High
Category
Privilege Escalation
Confidence
87% confidence
Finding

getCredential reads raw secret values from .env and returns them to the application, enabling the skill to inspect stored credentials. Even though status output masks previews, the underlying function still exposes full secret material to code paths, which is dangerous if the skill is extended, misused, or invoked by another component.

Content

Scanner excerpt · index.js (reported line 43)May include surrounding context.

js
}

/**
 * 读取 .env 文件中的凭据
 * @param {string} keyName - 环境变量名
 * @returns {string|null} - 凭据值或 null
 */

Credential Access

High
Category
Privilege Escalation
Confidence
93% confidence
Finding

saveCredential persists API keys in plaintext to .openclaw/.env, which is a sensitive credential-access and credential-storage behavior. In the skill context, this is more dangerous because the component is framed as a general assistant workflow helper, yet it handles secrets on disk without encryption, permission hardening, or safer storage primitives.

Content

Scanner excerpt · index.js (reported line 67)May include surrounding context.

js
}

/**
 * 保存凭据到 .env 文件
 * @param {string} keyName - 环境变量名
 * @param {string} keyValue - 凭据值
 * @returns {boolean} - 是否成功

Hidden Instructions

High
Category
Prompt Injection
Confidence
60% confidence
Finding

Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Content

Scanner excerpt · references/baidu-paddle.md (reported line 1)May include surrounding context.

md
# 闃块噷浜戠櫨鐐?Qwen Code 瀹樻柟璧勬簮

## 瀹樻柟缃戠珯
- **鐧剧偧鎺у埗鍙?*: https://bailian.console.aliyun.com

Hidden Instructions

High
Category
Prompt Injection
Confidence
60% confidence
Finding

Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Content

Scanner excerpt · references/claude-code.md (reported line 1)May include surrounding context.

md
# 闃块噷浜戠櫨鐐?Qwen Code 瀹樻柟璧勬簮

## 瀹樻柟缃戠珯
- **鐧剧偧鎺у埗鍙?*: https://bailian.console.aliyun.com

Hidden Instructions

High
Category
Prompt Injection
Confidence
60% confidence
Finding

Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Content

Scanner excerpt · references/codex-openai.md (reported line 1)May include surrounding context.

md
# 闃块噷浜戠櫨鐐?Qwen Code 瀹樻柟璧勬簮

## 瀹樻柟缃戠珯
- **鐧剧偧鎺у埗鍙?*: https://bailian.console.aliyun.com

Hidden Instructions

High
Category
Prompt Injection
Confidence
60% confidence
Finding

Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Content

Scanner excerpt · references/gemini-cli.md (reported line 1)May include surrounding context.

md
# 闃块噷浜戠櫨鐐?Qwen Code 瀹樻柟璧勬簮

## 瀹樻柟缃戠珯
- **鐧剧偧鎺у埗鍙?*: https://bailian.console.aliyun.com

Hidden Instructions

High
Category
Prompt Injection
Confidence
60% confidence
Finding

Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Content

Scanner excerpt · references/github-copilot.md (reported line 1)May include surrounding context.

md
# 闃块噷浜戠櫨鐐?Qwen Code 瀹樻柟璧勬簮

## 瀹樻柟缃戠珯
- **鐧剧偧鎺у埗鍙?*: https://bailian.console.aliyun.com

Hidden Instructions

High
Category
Prompt Injection
Confidence
60% confidence
Finding

Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Content

Scanner excerpt · references/openrouter.md (reported line 1)May include surrounding context.

md
# 闃块噷浜戠櫨鐐?Qwen Code 瀹樻柟璧勬簮

## 瀹樻柟缃戠珯
- **鐧剧偧鎺у埗鍙?*: https://bailian.console.aliyun.com

Hidden Instructions

High
Category
Prompt Injection
Confidence
60% confidence
Finding

Hidden instructions were detected in comments or invisible text. These could contain malicious directives. Manual review is recommended.

Content

Scanner excerpt · references/qwen-code.md (reported line 1)May include surrounding context.

md
# 闃块噷浜戠櫨鐐?Qwen Code 瀹樻柟璧勬簮

## 瀹樻柟缃戠珯
- **鐧剧偧鎺у埗鍙?*: https://bailian.console.aliyun.com

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The README describes the skill only as a general coding-plan assistant, but the actual documented behaviors include credential collection, storage in .openclaw/.env, registration guidance, pricing comparison, and purchase-related workflows. This mismatch reduces informed consent and makes it easier for users or invoking systems to authorize a skill without understanding that it handles secrets, which raises the risk of unsafe credential exposure or unintended privileged actions.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The credential-configuration flow is invoked by ordinary speech ('配置 GitHub Copilot') and immediately instructs the user to provide an API key for storage. Because this is a secret-handling capability, ambiguous activation is more dangerous in context: a routine conversation could transition into credential intake without strong user intent verification, leading to accidental disclosure of sensitive tokens to the agent or wrong skill.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The natural-language triggers are broad conversational phrases such as '帮我注册 Claude Code' and '配置 GitHub Copilot' without clear activation boundaries or confirmation requirements. In an agent environment, generic phrases can cause accidental invocation of sensitive flows like registration or credential handling when the user intended only to ask a question, increasing the chance of unintended collection or modification of secrets.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
84% confidence
Finding

The skill documents capability to read and manage API keys from .openclaw/.env, but the manifest declares no explicit tool scope or permissions. That mismatch weakens user consent and reviewability because a credential-handling skill can access sensitive local state without clearly advertising that need up front.

Content

No source excerpt is available for this finding.