Back to skill

Security audit

内容创业凭证管理器

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed credential manager, but it asks users to share and store high-value API keys and session cookies in ways that are too exposed for automatic approval.

Install only if you are comfortable with this skill handling credentials. Prefer entering secrets through a local secure prompt or OS keychain instead of chat or command-line arguments, avoid storing browser session cookies when possible, verify ~/.openclaw/credentials.json and its directory are private, and rotate any credentials that may have been pasted into chat or shell history.

Vulnerability Patterns
  • 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
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Findings (2)

T09 · Insecure Skill Coding Practices

Warning
Location
scripts/set_credential.py:45
Finding

Credentials Are Supplied Through Process Command-Line Arguments

Content
View full analysis
") print(f"平台: {', '.join(VALID_PLATFORMS)}") print("字段示例: api_key, access_key, secret_key, cookie, app_id, app_secret") print("\n示例:") print(" python3 set_credential.py tavily api_key tvly-xxxxx") print(" python3 set_credential.py stepfun api_key sk-xxxxx") print(" python3 set_credential.py wechat app_id wx_xxxxx") sys.exit(1) platform = sys.argv[1] field = sys.argv[2] value = sys.argv[3] ``` The documented invocation reinforces this behavior: ```bash python3 skills/content-credential-manager/scripts/set_credential.py ``` ### Technical Analysis The credential value is accepted directly through `sys.argv[3]`. Command-line arguments are not an appropriate transport for API keys, application secrets, or authenticated browser cookies. Depending on the operating system and process-isolation configuration, command-line arguments may be exposed through: - Process inspection tools such as `ps`. - Process metadata interfaces such as `/proc//cmdline`. - Shell command history. - Terminal session logging or audit tooling. - Process-monitoring and endpoint-management agents. - Agent execution logs that record commands before invocation. Masking the value after it has been processed does not protect it from these channels. The masking function only affects the script's final output, while the original value remains present in the process argument vector for the process lifetime. The risk is especially significant for `xiaohongshu.cookie`, because an authenticated session cookie may enable direct account impersonation withou ...[truncated 1737 chars]
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
scripts/set_credential.py:31
Finding

Claimed Credential File Permissions Are Not Enforced Reliably

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • 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 (17)

Tp4

High
Category
MCP Tool Poisoning
Confidence
92% confidence
Finding

The skill claims to track status, guide setup, and diagnose availability, but the documented behavior centers on collecting secrets from chat and writing them to a local file. This mismatch is dangerous because users may grant trust for a benign-seeming management function without realizing the skill performs sensitive credential persistence and handling.

Content

No source excerpt is available for this finding.

Credential Access

High
Category
Privilege Escalation
Confidence
88% confidence
Finding

The skill defines a centralized plaintext credential store at ~/.openclaw/credentials.json for multiple high-value services, including API secrets and session cookies. Consolidating sensitive credentials in a predictable file path raises the impact of any local compromise, accidental logging, overbroad file access, or unsafe script behavior.

Content

Scanner excerpt · SKILL.md (reported line 23)May include surrounding context.

凭证配置文件

所有凭证保存在 ~/.openclaw/credentials.json:

json
{

Ssd 3

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

The skill explicitly tells users to disclose full API secrets and complete Xiaohongshu session cookies in chat so the assistant can store them. In this context, the skill is a credential manager, which makes the issue more dangerous because the entire workflow normalizes transmission of highly sensitive authentication material through a less secure conversational channel.

Content

No source excerpt is available for this finding.

Credential Access

High
Category
Privilege Escalation
Confidence
95% confidence
Finding

This section instructs the user to provide a Tavily API key so the assistant can write it into the local credential file. The danger is not merely accessing credentials, but actively soliciting and persisting them through the agent, expanding exposure to chat logs, agent memory, and local storage.

Content

Scanner excerpt · SKILL.md (reported line 81)May include surrounding context.

md
1. 去 https://app.tavily.com 注册
2. Get API Key → 复制
3. 告诉我:"Tavily API Key 是 tvly-xxxxx"
4. 我帮你写入 `~/.openclaw/credentials.json`

#### 2. 阶跃星辰 TTS(5分钟)
1. 去 https://platform.stepfun.com 注册

Missing User Warnings

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The skill instructs users to paste API keys, app secrets, and full session cookies directly into chat, but it does not provide a strong warning that these values can grant account or API access if exposed. Session cookies are especially sensitive because they may allow direct account hijacking without needing a password or MFA.

Content

No source excerpt is available for this finding.

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 13)May include surrounding context.

md
import os
import sys

CREDENTIALS_FILE = os.path.expanduser("~/.openclaw/credentials.json")

# 技能凭证清单
CREDENTIALS_CHECKLIST = {

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 · scripts/check_status.py (reported line 10)May include surrounding context.

python
import os
import sys

CREDENTIALS_FILE = os.path.expanduser("~/.openclaw/credentials.json")

# 技能凭证清单
CREDENTIALS_CHECKLIST = {

Credential Access

High
Category
Privilege Escalation
Confidence
75% confidence
Finding

The script stores sensitive API credentials in a predictable plaintext file under the user's home directory. Although it attempts to restrict permissions via umask, the use of a local JSON file still increases exposure risk from local compromise, backups, misconfigured existing file permissions, or other processes reading secrets at rest; in this skill context, the file aggregates multiple platform credentials, which raises the blast radius.

Content

Scanner excerpt · scripts/set_credential.py (reported line 11)May include surrounding context.

python
import os
import sys

CREDENTIALS_FILE = os.path.expanduser("~/.openclaw/credentials.json")

VALID_PLATFORMS = ["tavily", "stepfun", "meitu", "xiaohongshu", "wechat"]
VALID_FIELDS = {

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The README tells users to pass secret credential values directly as command-line arguments to set_credential.py. On many systems, secrets provided this way can be exposed through shell history, process listings, terminal scrollback, or audit logs, which can leak API keys and cookies to other local users or later forensic collection.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
80% confidence
Finding

The skill describes reading and writing local credential files and invoking local scripts, but it does not declare any explicit tool scope or permission boundaries. That makes sensitive file access implicit rather than auditable, increasing the chance that an agent can handle secrets in ways the user did not clearly authorize.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The skill description, workflow, examples, and user-facing instructions are all presented in Chinese, including the expected phrases users should say. There is no indication that the user may choose another language or that the locale restriction is intentional for a region-specific tool.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The script’s docstrings and all user-facing print messages are written only in Chinese, with no option for the user to select another language. Under the policy, forcing a specific language without user opt-in is a natural-language policy violation unless the locale restriction is explicitly documented and justified.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

This file contains natural-language instructions and output that force a specific language for all users. Under the policy, language constraints should either be optional for the user or clearly justified as region-specific, which is not present here.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The trigger phrase '哪些技能可以用了' is broad natural language and can plausibly match routine status questions outside the narrow credential-management workflow. This can cause unintended skill activation, leading to confusing routing or disclosure of credential-status metadata when another skill or general assistant behavior would have been more appropriate.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The phrase '凭证配置好了吗' is ambiguous because it does not identify which credentials, which platform, or even that the request is about this content-creation skill set. In a multi-skill environment, this ambiguity increases the chance of accidental invocation and unnecessary exposure of configuration state across unrelated workflows.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The placeholder trigger '为什么XX技能不能用' is especially vague because 'XX技能' can stand in for nearly any skill name, greatly expanding activation scope. In context, this manager tracks installed skills and credential states, so an overbroad troubleshooting trigger could intercept many unrelated support queries and expose internal status or misroute users away from the correct specialized skill.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The trigger '开始内容创作前检查环境' is broad operational language that could overlap with many setup, publishing, or creative-workflow assistants. Because this skill specifically manages API credential readiness, the lack of explicit credential scoping raises the risk of unintended activation and inappropriate environment/status checks.

Content

No source excerpt is available for this finding.