Back to skill

Security audit

紫微识人管人报告生成

Security checks for vulnerabilities and agentic risk

Overview

This skill is not executable malware, but it asks agents to collect and retain sensitive employee profiling data with weak consent, privacy, and scope controls.

Review carefully before installing. Use only with explicit consent from every person whose birth or workplace data is processed, avoid submitting names or employment identifiers to external charting sites, disable MEMORY.md and default report archival unless there is an approved retention policy, and do not use these profiles as a sole basis for employment decisions.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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 (4)

T01 · Skill Instruction Hijacking

Error
Location
SKILL.md:15
Finding

User Request Scope Override Through Skill Instructions

Content
View full analysis
Remediation
View remediation

other

Error
Location
SKILL.md:26
Finding

External Disclosure of Sensitive Personal and Employment-Linked Data

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Error
Location
SKILL.md:181
Finding

Persistent Storage and Onward Sharing of Identifiable Personnel Profiles Without Safeguards

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
SKILL.md:14
Finding

Runtime Mutation of the Trusted Skill Instruction File

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (7)

Missing User Warnings

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

The skill instructs collection of highly sensitive personal data, including precise birth date/time, sex, and workplace relationship details, without any visible privacy notice, consent flow, or handling limitations. In a people-management context, this data can enable intrusive profiling, discrimination, or unauthorized workplace decision-making if processed without clear safeguards.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

The skill directs persistent storage of personnel reports and memory records without warning users about retention, access, or disclosure. Because these records concern employee traits, weaknesses, and management strategies, undisclosed persistence materially raises the risk of privacy violations, insider misuse, and harmful downstream employment decisions.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The skill is described as generating a Ziwei-based people-management report from chart input, but its documented workflow expands scope to collecting birth date/time, sex, organizational role, reporting lines, and recent incidents. That is a material increase in data collection beyond a simple report-generation claim, and it creates privacy and misuse risk because the extra personnel context is sensitive and directly tied to workplace evaluation.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The manifest presents the skill as report generation, but the instructions additionally require writing reports to local files, updating MEMORY.md, and potentially forwarding outputs for review. That changes the skill from transient analysis to persistent storage and distribution of employee profiles, increasing confidentiality, retention, and secondary-use risk without clear disclosure.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

Long-term archival of personnel reports and memory updates is not inherently necessary to produce a one-time Ziwei analysis report. Retaining detailed profiles creates an avoidable stockpile of sensitive employee data that can later be searched, reused, or leaked for purposes unrelated to the original request.

Content

No source excerpt is available for this finding.

Ssd 3

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

Instructions to write detailed personal reports into local archives and a memory file create a natural-language data retention and leakage risk even without executable code. Sensitive HR-style notes about individuals can be exposed through filesystem access, later prompt retrieval, backups, or accidental sharing, especially when the skill normalizes permanent accumulation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The entire report template is written in Chinese, beginning with the title on L001 and continuing throughout all section headings and field labels. This imposes a specific language/locale on generated output, and the file does not indicate that the user can choose another language or explicitly opt in to Chinese.

Content

No source excerpt is available for this finding.