Back to skill

Security audit

高等数学作业全生命周期管理套件

Security checks for vulnerabilities and agentic risk

Overview

The skill appears to be a legitimate calculus homework suite, but it handles student records and external AI/OCR services without enough privacy, data-retention, or installation safeguards.

Review this before installing in any real classroom or institution. Confirm where student work, images, grades, profiles, and analytics are stored; whether any data is sent to LLM, OCR, speech, or visualization providers; how consent, retention, deletion, access controls, and audit logging are handled; and pin or verify every external component before use. Avoid putting real API keys directly in command-line arguments.

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)

T08 · Insecure Dependencies

Warning
Location
README.md:34
Finding

Unpinned Global Installation of an External CLI Package

Content
View full analysis

Vulnerability Details

File Location: README.md:34
Vulnerability Type: Unpinned global third-party package installation
Risk Level: Medium

Vulnerable Code:

bash
npm install -g clawhub

Technical Analysis

The installation command retrieves the current registry-selected version of the clawhub package without an exact version, integrity hash, signature check, or documented provenance verification. The -g option installs the package globally and exposes its executable commands system-wide.

Because the package version is mutable from the user's perspective, the code executed during installation can differ from the code originally reviewed by the project author. npm packages may also define lifecycle scripts that execute during installation. This creates a supply-chain boundary in which compromise of the package, publisher account, or configured registry can result in execution of attacker-controlled code with the privileges of the user performing the installation.

Attack Path

  1. An attacker compromises the clawhub publisher account, package, or package distribution channel.
  2. The attacker publishes a malicious or otherwise altered version under the expected package name.
  3. A user follows the documented npm install -g clawhub command.
  4. npm resolves and downloads the mutable package version without project-supplied integrity verification.
  5. Any malicious installation lifecycle code executes with the invoking user's privileges.
  6. The globally installed executable remains available for later commands and can perform additional attacker-controlled actions when invoked.

Impact Assessment

Successful exploitation can provide code execution under the account that runs npm. The affected scope may include that user's files, credentials, environment variables, and accessible development resources. If installation is performed with elevated privileges, the impact can extend to system ...[truncated 205 chars]

Remediation
View remediation

Remediation Suggestions

  • Pin the CLI to an exact, reviewed version rather than allowing registry-selected updates.
  • Document the authoritative package registry and verified publisher identity.
  • Use npm lockfile integrity data or equivalent cryptographic provenance controls.
  • Verify package signatures or registry provenance attestations before installation.
  • Prefer a project-local development dependency invoked through npx --no-install or an npm script instead of a global installation.
  • Run installation as a non-privileged user and prohibit sudo npm install -g.
  • Review the package and its transitive dependencies, including all npm lifecycle scripts, before recommending installation.

T08 · Insecure Dependencies

Warning
Location
package.json:27
Finding

Mutable and Unaudited External Components Supply Core Functionality

Content
View full analysis

Vulnerability Details

File Location: package.json:27-32; related installation instructions at SKILL.md:92-95
Vulnerability Type: Mutable third-party dependency resolution
Risk Level: Medium

Vulnerable Code in package.json:

json
"dependencies": {
  "calculus-homework-assignment": "^1.0.0",
  "calculus-intelligent-grading": "^1.0.0",
  "calculus-learning-analytics": "^1.0.0",
  "calculus-error-analyzer": "^1.0.0"
},

Related Installation Commands in SKILL.md:

bash
clawhub install calculus-homework-assignment
clawhub install calculus-intelligent-grading
clawhub install calculus-learning-analytics
clawhub install calculus-error-analyzer

Technical Analysis

All substantive functionality is delegated to four external components that are not included in the audited artifact. The npm dependency declarations use caret ranges, permitting compatible releases newer than 1.0.0. The ClawHub installation commands identify components only by name and do not specify an immutable version, content digest, signature, or trusted publisher identity.

Consequently, installing this package or following its documentation may retrieve code that was not part of this audit. The supplied directory contains no lockfile or vendored dependency source through which the resolved artifacts and integrity hashes could be verified. This creates exposure to compromised publisher accounts, malicious updates, dependency confusion, registry substitution, and unexpected behavior introduced by later releases.

Attack Path

  1. An attacker compromises a named dependency, its publisher account, or an applicable package registry.
  2. The attacker publishes a malicious version accepted by the caret range, or causes a name-only ClawHub installation to resolve to attacker-controlled content.
  3. A user installs the suite or follows the documented component installation commands.
  4. The package manager res ...[truncated 843 chars]
Remediation
View remediation

Remediation Suggestions

  • Pin every dependency and ClawHub component to an exact reviewed version.
  • Commit a lockfile containing resolved sources and integrity hashes.
  • Require cryptographic signatures or provenance attestations and verify publisher identities.
  • Explicitly configure approved registries and disable fallback to untrusted or public sources where dependency confusion is possible.
  • Audit each external component and its transitive dependencies before installation.
  • Consider vendoring security-critical components or recording immutable artifact digests.
  • Add automated dependency scanning, update review, and allowlist enforcement to the release pipeline.
  • Run installed components with least privilege and restrict filesystem, credential, database, and network access.

T09 · Insecure Skill Coding Practices

Warning
Location
README.md:54
Finding

API Credentials Passed Through Command-Line Arguments

Content
View full analysis

Vulnerability Details

File Location: README.md:54 and README.md:238
Vulnerability Type: Insecure credential handling in documented commands
Risk Level: Medium

Vulnerable Code:

bash
clawhub publish --api-key YOUR_API_KEY
bash
clawhub login --username your_username --api-key your_key

Technical Analysis

The examples instruct users to replace placeholders with API credentials directly in command-line arguments. Command-line secrets may be recorded in shell history, terminal session logs, CI/CD logs, troubleshooting output, or process-monitoring systems. Depending on the operating system and process configuration, other local users or processes may also be able to inspect process arguments while the command is running.

Although the repository contains placeholders rather than actual credentials, following the documented pattern with a real key creates a credential-disclosure risk.

Attack Path

  1. A user substitutes a real ClawHub API key into one of the documented commands.
  2. The shell stores the complete command in its history, or an automation platform records it in build logs.
  3. Alternatively, a local process observes the command arguments while the command is running.
  4. An attacker with access to the history file, logs, terminal records, or process metadata retrieves the key.
  5. The attacker uses the exposed credential against ClawHub within the permissions assigned to that key.

Impact Assessment

The attacker can obtain the privileges associated with the compromised API key. Depending on server-side authorization, this may include authenticating as the user, publishing or modifying skill packages, accessing account resources, or distributing altered releases. This could also enable a downstream supply-chain attack against users who trust packages published by the affected account. The exact scope is limited by the permissions, expiry, and revocation controls of t ...[truncated 22 chars]

Remediation
View remediation

Remediation Suggestions

  • Remove examples that place API keys directly in command-line arguments.
  • Use an interactive login flow that reads secrets from standard input without echoing them.
  • Alternatively, use a protected environment variable or an operating-system credential store, provided the CLI prevents accidental logging.
  • Ensure CI/CD systems inject credentials through their secret-management facilities and mask them in logs.
  • Configure short-lived, narrowly scoped tokens rather than long-lived account-wide keys.
  • Document immediate token rotation and revocation procedures for suspected exposure.
  • Add automated secret scanning and log-redaction controls.
  • If command-line credential support must remain available, clearly warn users about shell history and process-list exposure.
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (14)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The README describes processing student analytics, class dashboards, OCR/image submissions, and learning profiles, but provides no privacy notice, consent requirements, retention rules, or safeguards for handling educational data. In an education context, this is sensitive personal data, and documenting collection/processing workflows without data-handling constraints increases the risk of misuse, overcollection, and noncompliance with privacy obligations.

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 process student submissions, personal ability profiles, and class analytics, but the documentation provides no privacy, consent, retention, or access-control warnings. In an education context, this can lead users to handle regulated or sensitive student data without safeguards, increasing the risk of unauthorized disclosure or noncompliant processing.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The grading example shows image-based student work being uploaded or processed, but there is no warning that images may contain personally identifiable information, handwritten signatures, or other sensitive content. Without explicit handling guidance, operators may transmit student work to services or logs insecurely, exposing private educational records.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The environment variables configure external LLM and OCR services, implying that student submissions and analytics data may be sent to third parties, yet the documentation does not disclose this data flow or its privacy implications. This is more dangerous in the skill context because the data concerns students and class performance, which may be protected by law or institutional policy.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
82% confidence
Finding

The skill documentation is entirely in Chinese and does not indicate that users may choose another language or locale. Under the stated policy, forcing a specific language without user opt-in can be a natural-language policy violation unless the locale restriction is explicitly justified.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
84% confidence
Finding

The natural-language content of the skill file is exclusively Chinese, which can amount to a language/locale constraint when no opt-in or justification is given. SQP-3 requires flagging language policy issues unless the skill offers a language choice or clearly documents a justified locale restriction.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
93% confidence
Finding

This manifest contains natural-language metadata only in Chinese, which can constitute a language/locale policy issue when no user opt-in or justification is provided. The file gives no indication that the skill is region-specific or that users can choose another language.

Content

No source excerpt is available for this finding.

Unpinned Dependencies

Low
Category
Supply Chain
Confidence
90% confidence
Finding

Using a caret version range allows newer compatible releases to be installed automatically, which can introduce a compromised or vulnerable dependency through the supply chain without explicit review. In a skill package, this risk is amplified because multiple external educational components are pulled in and may execute in the host environment.

Content

Scanner excerpt · package.json (reported line 28)May include surrounding context.

json
"openclaw": ">=2026.4.0"
  },
  "dependencies": {
    "calculus-homework-assignment": "^1.0.0",
    "calculus-intelligent-grading": "^1.0.0",
    "calculus-learning-analytics": "^1.0.0",
    "calculus-error-analyzer": "^1.0.0"

Unpinned Dependencies

Low
Category
Supply Chain
Confidence
90% confidence
Finding

Using a caret version range allows newer compatible releases to be installed automatically, which can introduce a compromised or vulnerable dependency through the supply chain without explicit review. In a skill package, this risk is amplified because multiple external educational components are pulled in and may execute in the host environment.

Content

Scanner excerpt · package.json (reported line 29)May include surrounding context.

json
},
  "dependencies": {
    "calculus-homework-assignment": "^1.0.0",
    "calculus-intelligent-grading": "^1.0.0",
    "calculus-learning-analytics": "^1.0.0",
    "calculus-error-analyzer": "^1.0.0"
  },

Unpinned Dependencies

Low
Category
Supply Chain
Confidence
90% confidence
Finding

Using a caret version range allows newer compatible releases to be installed automatically, which can introduce a compromised or vulnerable dependency through the supply chain without explicit review. In a skill package, this risk is amplified because multiple external educational components are pulled in and may execute in the host environment.

Content

Scanner excerpt · package.json (reported line 30)May include surrounding context.

json
"dependencies": {
    "calculus-homework-assignment": "^1.0.0",
    "calculus-intelligent-grading": "^1.0.0",
    "calculus-learning-analytics": "^1.0.0",
    "calculus-error-analyzer": "^1.0.0"
  },
  "peerDependencies": {

Unpinned Dependencies

Low
Category
Supply Chain
Confidence
90% confidence
Finding

Using a caret version range allows newer compatible releases to be installed automatically, which can introduce a compromised or vulnerable dependency through the supply chain without explicit review. In a skill package, this risk is amplified because multiple external educational components are pulled in and may execute in the host environment.

Content

Scanner excerpt · package.json (reported line 31)May include surrounding context.

json
"calculus-homework-assignment": "^1.0.0",
    "calculus-intelligent-grading": "^1.0.0",
    "calculus-learning-analytics": "^1.0.0",
    "calculus-error-analyzer": "^1.0.0"
  },
  "peerDependencies": {
    "calculus-concept-visualizer": "^1.0.0",

Unpinned Dependencies

Low
Category
Supply Chain
Confidence
82% confidence
Finding

The peer dependency is also specified with a caret range, so consumers may resolve to different package versions over time, increasing supply-chain uncertainty. This is somewhat less direct than a runtime dependency because peer dependencies are installed and controlled by the consuming environment, but it can still expose users to unreviewed versions.

Content

Scanner excerpt · package.json (reported line 34)May include surrounding context.

json
"calculus-error-analyzer": "^1.0.0"
  },
  "peerDependencies": {
    "calculus-concept-visualizer": "^1.0.0",
    "question-type-generator": "^1.0.0",
    "math-edu-assistant": "^1.0.0"
  },

Unpinned Dependencies

Low
Category
Supply Chain
Confidence
82% confidence
Finding

The peer dependency is also specified with a caret range, so consumers may resolve to different package versions over time, increasing supply-chain uncertainty. This is somewhat less direct than a runtime dependency because peer dependencies are installed and controlled by the consuming environment, but it can still expose users to unreviewed versions.

Content

Scanner excerpt · package.json (reported line 35)May include surrounding context.

json
},
  "peerDependencies": {
    "calculus-concept-visualizer": "^1.0.0",
    "question-type-generator": "^1.0.0",
    "math-edu-assistant": "^1.0.0"
  },
  "scripts": {

Unpinned Dependencies

Low
Category
Supply Chain
Confidence
82% confidence
Finding

The peer dependency is also specified with a caret range, so consumers may resolve to different package versions over time, increasing supply-chain uncertainty. This is somewhat less direct than a runtime dependency because peer dependencies are installed and controlled by the consuming environment, but it can still expose users to unreviewed versions.

Content

Scanner excerpt · package.json (reported line 36)May include surrounding context.

json
"peerDependencies": {
    "calculus-concept-visualizer": "^1.0.0",
    "question-type-generator": "^1.0.0",
    "math-edu-assistant": "^1.0.0"
  },
  "scripts": {
    "install": "echo 'Installing calculus homework suite...'",