Back to skill

Security audit

POLICY-MANAGER

Security checks for vulnerabilities and agentic risk

Overview

This skill is for insurance policy management, but it handles sensitive records with weak file boundaries, no implemented authorization checks, and under-disclosed network and storage behavior.

Review before installing. This appears to be a purpose-built insurance policy helper, not an obvious malicious package, but it should only run in a controlled environment with trusted callers and limited filesystem access. Add task number validation, authorization, update allowlists, schema validation, safe writes, explicit privacy notices, and clear opt-in for any remote POLICY_API_URL use before using it with real customer data.

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 (4)

T05 · Unauthorized Access and Privilege Escalation

Error
Location
scripts/policy-manager.js:114
Finding

Path Traversal Enables Access to JSON Files Outside the Policy Directory

Content
View full analysis
Remediation
View remediation

T05 · Unauthorized Access and Privilege Escalation

Error
Location
scripts/policy-manager.js:119
Finding

Policy Records Can Be Read or Modified Without Authorization Checks

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
scripts/policy-manager.js:145
Finding

Caller-Controlled Update Type Permits Unrestricted Top-Level Policy Modification

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
scripts/policy-manager.js:145
Finding

Missing File Locking and Non-Atomic Writes Permit Lost Updates and Policy Corruption

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (9)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

This markdown file enumerates many personally sensitive fields such as certNo, mobile, bankAccNo, taxpayerNo, payerCertNo, and anti-money-laundering data, but it does not include any warning or disclosure about handling personal, financial, or regulated data. Under the markdown-file criteria for SQP-2, documentation that describes behavior or data affecting privacy should warn users about those implications.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
95% confidence
Finding

The skill instructs the agent to invoke multiple CLI-style tools and handle policy files, but it does not declare any explicit tool scope or permissions boundaries. In a skill that processes sensitive insurance data, missing tool restrictions increases the risk of unintended tool access, environment exposure, or broader-than-necessary execution if the runtime grants default capabilities.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill description covers creating and updating policy JSON files and parsing uploaded materials, but it gives no warning that these workflows involve highly sensitive personal and financial data such as ID numbers, phone numbers, bank card details, and insurance records. Without upfront warning and consent-oriented language, users may overshare sensitive data or operators may handle it without appropriate privacy safeguards.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The operational sections direct the agent to write and update policy JSON files containing PII, financial information, and document-derived data, yet there is no user warning or confirmation step before persistence. This creates privacy and compliance risk because sensitive information may be stored, updated, or propagated across tools without explicit notice, minimization, or confirmation.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The skill’s stated purpose is local JSON policy management, but it also supports outbound HTTP(S) calls to a configurable external API to fetch policy templates. That creates an unexpected data flow and trust boundary expansion: insurance/product identifiers are sent off-host, and if POLICY_API_URL is misconfigured or attacker-controlled, the skill may contact an untrusted service and ingest untrusted template data.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The remote API request includes insuranceCode and productCode in the query string without clear disclosure to the user. Even if these are not full policy records, they are business-sensitive identifiers, and transmitting them to a configurable external endpoint can leak workflow or customer context and expose them in logs, proxies, and monitoring systems.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The skill writes policy data containing potentially sensitive insurance and personal information to local disk without any user disclosure, consent, or storage safeguards. In the context of insurance workflows, silent persistence increases privacy and compliance risk, especially on shared machines or developer workstations.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The update path overwrites existing policy files in place without disclosure or safeguards such as versioning, integrity checks, or confirmation. For sensitive policy records, this can lead to silent data loss, unauthorized modification persistence, or tampering that is hard to detect.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Low
Category
Not specified by scanner
Confidence
80% confidence
Finding

The stated purpose is managing policy JSON files through create/read/update flows. Accessing an unrelated mock API dataset at a fixed external path is an additional capability that is not clearly justified by the manifest's narrow policy-management scope, especially as a production behavior fallback.

Content

No source excerpt is available for this finding.