Back to skill

Security audit

Ai Intelligent Employee Onboarding

Security checks for vulnerabilities and agentic risk

Overview

The skill describes a plausible HR onboarding tool, but it asks users to run unreviewed mutable remote code for sensitive employee workflows without clear safeguards.

Review this carefully before installing. Use an isolated environment, inspect the referenced GitHub repository and dependencies yourself, pin to a known commit, and do not connect it to HR, device-provisioning, training, benefits, or notification systems until data handling, approvals, and rollback procedures are clear.

Vulnerability Patterns
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • 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
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Findings (1)

T03 · Remote Payload Retrieval and Execution

Error
Location
SKILL.md:40
Finding
Unpinned Remote Payload Retrieval and Execution## Vulnerability Details **File Location**: `SKILL.md`, lines 40–43 **Vulnerability Type**: Remote payload retrieval and execution **Risk Level**: High ```bash git clone https://github.com/openclaw-skills/ai-intelligent-employee-onboarding cd ai-intelligent-employee-onboarding pip install -r requirements.txt python app.py ``` ### Technical Analysis The installation instructions retrieve a mutable external Git repository without pinning it to a reviewed commit or immutable release. They subsequently install dependencies from the retrieved `requirements.txt` and execute the retrieved `app.py`. The remote source code and dependency manifest are not included in the audited artifact, which contains only `SKILL.md` and `skill.json`. Consequently, the effective payload cannot be verified during this audit and may change after publication or review. Both Python package installation hooks and `app.py` can execute arbitrary code with the privileges of the user following these instructions. Package provenance is also unclear because `skill.json` identifies the author as `yang1002378395-cmyk`, while `SKILL.md` attributes the project to “OpenClaw Skills Team.” This discrepancy does not independently prove malicious behavior, but it limits confidence in the external repository's ownership. ### Attack Path 1. An attacker compromises the referenced repository, gains control of its publishing account, or otherwise causes malicious content to be added to its default branch. 2. The attacker modifies `app.py`, `requirements.txt`, or a referenced dependency to include an arbitrary payload. 3. A user follows the documented installation procedure and clones the current default branch without an immutable revision pin. 4. `pip install -r requirements.txt` may execute malicious package installation or build hooks. 5. `python app.py` directly executes the remotely controlled application payload. 6. The payload operates with the invoking user ...[truncated 687 chars]
Remediation
## Remediation Suggestions 1. Include the complete, auditable application source and dependency manifests directly in the Skill package. 2. If external retrieval is unavoidable, pin the repository to a reviewed full commit SHA rather than cloning a mutable default branch. 3. Verify downloaded content using cryptographic hashes or signed releases before installation or execution. 4. Pin every Python dependency to an exact reviewed version and require package hashes, such as through `pip install --require-hashes`. 5. Generate and review a software bill of materials and use automated dependency vulnerability scanning. 6. Publish verifiable repository ownership and reconcile the author discrepancy between `skill.json` and `SKILL.md`. 7. Execute installation and startup in a least-privileged, isolated environment without unnecessary credentials, sensitive mounts, or host access. 8. Replace direct execution instructions with a reproducible build process based on immutable, reviewed artifacts.
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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (2)

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The skill describes collecting employee information and triggering automated onboarding actions such as device requests, training scheduling, and gift distribution, but it provides no warning about handling sensitive personal data, approval requirements, or the consequences of automation. In an HR context, missing disclosure and safeguards can lead to over-collection, unauthorized actions, privacy violations, or accidental provisioning based on incomplete or incorrect inputs.

Natural-Language Policy Violations

Low
Confidence
91% confidence
Finding
The description and user-facing documentation are entirely in Chinese, with no indication that the skill is intentionally limited to Chinese-speaking users or a China-specific deployment. This can violate language/locale policy where users should be given a choice or the restriction should be explicitly justified.

Static analysis

No suspicious patterns detected.