Back to skill

Security audit

GitHub 人才猎手 (GitHub Talent Hunter)

Security checks for vulnerabilities and agentic risk

Overview

The skill mostly does what it says, but it asks users for broader GitHub token access than its public search function appears to need and gives unsafe token-storage guidance.

Review before installing. Use a dedicated short-lived GitHub token with the minimum scopes that actually work for your use case, avoid granting read:org unless you verify it is required, and prefer a temporary environment variable or secret manager over a plaintext .env file. Treat saved candidate results as personal data and follow privacy, anti-spam, and employment-law requirements before contacting anyone.

Vulnerability Patterns
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
  • 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)

T05 · Unauthorized Access and Privilege Escalation

Warning
Location
SECURITY.md:52
Finding

GitHub Token Is Requested with Permissions Beyond the Public Search Requirement

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
SECURITY.md:57
Finding

Persistent Plaintext Token Guidance Relies on a Missing Git Ignore Rule and Does Not Restrict File Permissions

Content
View full analysis
~/.openclaw/workspace/skills/ai-talent-hunter/.env grep ".env" ~/.openclaw/workspace/skills/ai-talent-hunter/.gitignore ``` A similar command appears in `SKILL.md`: ```bash echo "GITHUB_TOKEN=your_token_here" > .env ``` ### Technical Analysis The documentation recommends writing the GitHub token to a persistent plaintext `.env` file. The command does not establish restrictive permissions before or after writing the credential. Its resulting accessibility therefore depends on the user's umask, directory permissions, backup configuration, and local environment. The documentation repeatedly states that `.env` is already excluded by `.gitignore`, but the audited project directory structure contains no `.gitignore` file. Consequently, users may believe the credential is protected from version-control inclusion when that protection is not present in the reviewed artifact. There is also a functional inconsistency: `scripts/github_search.py` reads only the process environment through: ```python self.token = token or os.getenv("GITHUB_TOKEN") ``` No `.env` loader was identified. Unless OpenClaw or another external launcher loads `.env`, merely creating the file does not make the token available to the script. This may encourage users to duplicate the token across files and shell configuration while troubleshooting. The token is not hard-coded in the repository, and the reviewed script does not read arbitrary credential paths. The vulnerability is specifically the unsafe and misleading credential-setup guidance. ### Attack Path 1. A user replaces the placeholder with a real GitHub token and runs the documented `echo` command. 2. The tok ...[truncated 1229 chars]
Remediation
View remediation
.env ``` 5. Avoid passing real tokens directly as command-line arguments because process listings and shell history may expose them. 6. Either explicitly load `.env` using a pinned and reviewed mechanism or remove `.env` instructions and document only the supported environment-variable workflow. 7. Set restrictive permissions on the containing data directory and candidate database because candidate emails and profile data are also persisted locally. 8. Add documentation for token expiration, rotation, revocation, local backups, and incident response. 9. Add a startup check that warns when `.env` is group-readable, world-readable, tracked by Git, or absent from ignore rules. ]]>
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Findings (31)

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 68)May include surrounding context.

md
1. **需要外部 API 凭证(GitHub Token)**
   - 本 skill 使用 GitHub API 搜索开发者
   - 需要用户自行申请并配置 GitHub Personal Access Token
   - Token 存储在本地 `.env` 文件中(不会上传到 git)

2. **包含 Prompt 模板**

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 · SECURITY.md (reported line 11)May include surrounding context.

md
1. **需要外部 API 凭证(GitHub Token)**
   - 本 skill 使用 GitHub API 搜索开发者
   - 需要用户自行申请并配置 GitHub Personal Access Token
   - Token 存储在本地 `.env` 文件中(不会上传到 git)

2. **包含 Prompt 模板**

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 13)May include surrounding context.

md
1. **需要外部 API 凭证(GitHub Token)**
   - 本 skill 使用 GitHub API 搜索开发者
   - 需要用户自行申请并配置 GitHub Personal Access Token
   - Token 存储在本地 `.env` 文件中(不会上传到 git)

2. **包含 Prompt 模板**

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 138)May include surrounding context.

md
1. **需要外部 API 凭证(GitHub Token)**
   - 本 skill 使用 GitHub API 搜索开发者
   - 需要用户自行申请并配置 GitHub Personal Access Token
   - Token 存储在本地 `.env` 文件中(不会上传到 git)

2. **包含 Prompt 模板**

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 524)May include surrounding context.

md
1. **需要外部 API 凭证(GitHub Token)**
   - 本 skill 使用 GitHub API 搜索开发者
   - 需要用户自行申请并配置 GitHub Personal Access Token
   - Token 存储在本地 `.env` 文件中(不会上传到 git)

2. **包含 Prompt 模板**

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/github_search.py (reported line 1206)May include surrounding context.

python
1. **需要外部 API 凭证(GitHub Token)**
   - 本 skill 使用 GitHub API 搜索开发者
   - 需要用户自行申请并配置 GitHub Personal Access Token
   - Token 存储在本地 `.env` 文件中(不会上传到 git)

2. **包含 Prompt 模板**

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 · SECURITY.md (reported line 59)May include surrounding context.

  1. 配置 Token
    bash
    # 方式1: 使用 .env 文件(推荐,持久化)
    echo "GITHUB_TOKEN=your_token_here" > ~/.openclaw/workspace/skills/ai-talent-hunter/.env
    
    # 方式2: 临时环境变量(不持久化)
    

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 · SECURITY.md (reported line 60)May include surrounding context.

  1. 配置 Token
    bash
    # 方式1: 使用 .env 文件(推荐,持久化)
    echo "GITHUB_TOKEN=your_token_here" > ~/.openclaw/workspace/skills/ai-talent-hunter/.env
    
    # 方式2: 临时环境变量(不持久化)
    

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 · SECURITY.md (reported line 68)May include surrounding context.

  1. 配置 Token
    bash
    # 方式1: 使用 .env 文件(推荐,持久化)
    echo "GITHUB_TOKEN=your_token_here" > ~/.openclaw/workspace/skills/ai-talent-hunter/.env
    
    # 方式2: 临时环境变量(不持久化)
    

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 · SKILL.md (reported line 535)May include surrounding context.

  1. 配置 Token
    bash
    # 方式1: 使用 .env 文件(推荐,持久化)
    echo "GITHUB_TOKEN=your_token_here" > ~/.openclaw/workspace/skills/ai-talent-hunter/.env
    
    # 方式2: 临时环境变量(不持久化)
    

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 · SKILL.md (reported line 536)May include surrounding context.

  1. 配置 Token
    bash
    # 方式1: 使用 .env 文件(推荐,持久化)
    echo "GITHUB_TOKEN=your_token_here" > ~/.openclaw/workspace/skills/ai-talent-hunter/.env
    
    # 方式2: 临时环境变量(不持久化)
    

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 · SECURITY.md (reported line 69)May include surrounding context.

  1. 验证 .gitignore
    bash
    # 确认 .env 在 .gitignore 中
    grep ".env" ~/.openclaw/workspace/skills/ai-talent-hunter/.gitignore
    

Token 安全红线

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 · prompts/jd_parser_prompt.md (reported line 155)May include surrounding context.

md
# Constraints (约束条件)

1. **输出必须是纯 JSON**,不包含任何 Markdown 代码块标记(`​`​`​`json`)
2. **search_queries 必须生成 2-3 个**,从不同维度覆盖(语言、生态、关键词组合)
3. **必须包含活跃度过滤**:`pushed:>2025-01-01`(不要加 stars 限制,冷门技术可能没有高星仓库)
4. **🚨 一条 search_query 中最多只能出现一个 `language:` 标签**。GitHub API 不支持同时搜索多种语言。如果需要搜多种语言,请拆分成多条 query。

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

This markdown file describes searching GitHub developers, building candidate profiles, ranking contactability, and generating outreach, but its warning section only covers public APIs, no auto-sending, and no private-data scraping. It does not clearly disclose the privacy and outreach implications for users handling personal profile data, which falls under markdown-file user warnings about behaviors affecting privacy.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

该文件是 markdown,适用 SQP-1。触发场景中的“帮我找几个…”“想找个…”“给我 5 个候选人”等表述非常接近日常对话,且文档未给出明确的边界或反例来说明哪些招聘相关请求才应触发该 skill,容易与一般性的找人/咨询表达重叠。

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

SQP-3 适用于所有文件类型。文档明确要求‘不要照搬英文 bio(要翻译/提炼为中文)’,并在多处规定展示内容需翻译为中文,但没有说明这是用户可选项,也没有将该 skill限定为仅面向中文招聘场景,构成语言/locale 强制策略问题。

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

虽然‘商业语言’本身不是问题,但结合上下文中的中文示例和多处中文翻译要求,这里形成了固定输出风格与语言的约束,却没有提供用户选择或明确说明仅适用于中文招聘场景。根据 SQP-3,这属于缺少语言/locale opt-in 的自然语言政策问题。

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The entire instruction set, role, tasking, examples, and required reasoning/output conventions are written to constrain the interaction to Chinese, with no opt-in or alternative language path. Per the policy, forcing a specific language or locale without user choice is a natural-language policy violation unless clearly justified as region-specific, which is not documented here.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The prompt's 'correct practice' section says to force both pushed:>2025-01-01 and stars:>5, but the formal constraints later explicitly state pushed:>2025-01-01 must be included and stars limits must not be added. This is an active contradiction in the skill's own documentation that can cause downstream behavior to diverge depending on which section the model follows.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The skill is explicitly designed to generate personalized recruiting outreach using a candidate’s background, projects, employer history, and other profile signals, but it provides no guardrails about lawful collection, consent, minimization, or appropriate handling of personal data. This can enable privacy-invasive profiling and targeted unsolicited contact at scale, especially when users are encouraged to reference detailed personal context to increase response rates.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

文件标题及全文内容均以中文编写,但没有说明这是面向特定中文地区用户的限定文档,也未向用户提供语言或地区选择。根据规则,未经用户选择而强制单一语言/locale 的自然语言约束属于应报告的组织政策风险。

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

This Python file contains user-facing natural-language descriptions, docstrings, and CLI messages entirely in Chinese, including the module title and command descriptions. Under the policy for natural-language violations, forcing a specific language without user opt-in is in scope, and the file does not indicate that the skill is region-specific or that another language can be selected.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The file's docstrings, CLI usage comments, runtime logs, and labels are predominantly in Chinese, with no indication that the user can select another language or that the script is intentionally region-specific. This can violate language/locale policy because it effectively forces one locale for normal operation and troubleshooting.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · scripts/github_search.py (reported line 53)May include surrounding context.

python
"Authorization": f"Bearer {self.token}",
            "Content-Type": "application/json"
        }
        self.endpoint = "https://api.github.com/graphql"
        self._last_request_time = 0.0

    MAX_RETRIES = 2  # 502/503 重试次数

External Transmission

Medium
Category
Data Exfiltration
Confidence
80% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · scripts/github_search.py (reported line 76)May include surrounding context.

python
time.sleep(self.REQUEST_INTERVAL - elapsed)

            try:
                response = requests.post(
                    self.endpoint,
                    headers=self.headers,
                    json=payload,