Back to skill

Security audit

理工科论文写作

Security checks for vulnerabilities and agentic risk

Overview

The skill itself is mostly transparent and local, but its README includes an unsafe optional manual install command from an unverified placeholder GitHub repository.

Prefer installing through ClawHub or a verified, pinned release. Do not use the README's placeholder git clone command unless the repository owner, commit, and contents are independently verified. Review generated academic content carefully, cite sources, and replace any simulated data before use.

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
README.md:42
Finding

Unverified Mutable Repository Used for Manual Skill Installation

Content
View full analysis

Vulnerability Details

File Location: README.md, line 42
Vulnerability Type: Unverified third-party supply-chain source
Risk Level: Medium

Vulnerable Code:

bash
git clone https://github.com/yourusername/zeelin-academic-paper.git ~/.openclaw/workspace/skills/zeelin-academic-paper

Technical Analysis

The documented manual installation command clones a repository directly into the active OpenClaw Skill directory. The repository URL contains the placeholder owner yourusername, does not correspond to a verified canonical source in the reviewed package, and is not pinned to a release tag or immutable commit. The command also performs no signature, checksum, or provenance verification.

A Git repository's default branch is mutable. Consequently, the content installed by this command can differ from the content that was audited. If the placeholder repository exists or is later registered by an attacker, or if the referenced repository is compromised, a user following the instructions could install attacker-controlled Skill instructions or executable files.

The reviewed package itself contains no malicious executable code, remote payload retrieval behavior, credential access, persistence mechanism, or data-exfiltration logic. The risk arises specifically from the unsafe manual installation instruction.

Attack Path

  1. An attacker creates, acquires, or compromises the repository identified by the documented URL.
  2. The attacker publishes a modified Skill package containing hostile Agent instructions, scripts, or configuration.
  3. A user follows the README and executes the provided git clone command.
  4. Git retrieves the mutable repository without validating its identity, release version, signature, or content hash.
  5. The attacker-controlled files are written directly to ~/.openclaw/workspace/skills/zeelin-academic-paper.
  6. OpenClaw discovers or loads the installed Skill.
  7. Malicious ins ...[truncated 901 chars]
Remediation
View remediation

Remediation Suggestions

  1. Replace the placeholder repository URL with the verified canonical repository controlled by the documented project owner.
  2. Pin installation to an immutable, reviewed commit hash rather than cloning a mutable default branch.
  3. Publish signed releases and verify the release signature before installation.
  4. Publish a SHA-256 checksum for each release archive and require checksum verification in the installation procedure.
  5. Download and validate the package in a staging directory before moving it into the active OpenClaw Skill directory.
  6. Configure the runtime to require explicit approval for newly declared capabilities and to deny script execution unless necessary.
  7. Keep the README version, Skill manifest version, repository release, and integrity metadata synchronized.
  8. Remove the manual clone instruction until a valid and verifiable canonical repository is available.
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (6)

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
83% confidence
Finding

SQP-3 applies to all file types and covers language or locale policy violations. This document presents all policy and safety information only in Chinese, and L112 states that use implies the user has read and agreed, which can exclude users who do not read Chinese and does not offer an alternative language or explicit locale justification.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The example trigger "帮我写一篇关于…的论文" is a broad natural-language invocation that can cause the skill to activate for generic requests to write an academic paper, without clear scoping or consent checks. In this context, that increases the risk of over-automation of academic writing and misuse for plagiarism or submission of AI-generated work presented as original scholarship.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The supported-scenarios section explicitly lists '中文论文写作', which indicates the skill is constrained to a specific language. The document does not offer users a language choice or frame this as an opt-in preference, which can violate language/locale policy requirements.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The file states the required language style as academic and concise, and the overall skill content and resources are explicitly Chinese-focused, but it does not offer users a choice of output language. This creates a locale/language policy concern because the skill effectively forces a specific language without user opt-in or a clearly documented regional justification.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The natural-language instruction explicitly says to produce a complete Chinese STEM paper workflow. This imposes a specific language requirement without offering the user a language choice or documenting a justified locale restriction, which matches the language/locale policy violation criteria.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The instruction "仅返回中文题目" forces a specific language for outputs. Under the policy, locale or language constraints should not be imposed unless the user is given a choice or the restriction is clearly justified as region-specific.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.