ClawHub
Back to skill

Security audit

subscription-churn-analysis

Security checks for vulnerabilities and agentic risk

Overview

The bundled workflow is local and user-controlled, but its broken churn scoring should not guide business decisions.

Do not rely on the rankings until scoring, account identifiers, missing-data handling, and grouping are corrected and validated. Use account IDs and review reports before sharing. The optional unsigned Windows app is unnecessary for this workflow and needs a separate assessment.

A.I.G

Vulnerability Patterns
  • 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
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns
  • 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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The code's intended theme matches subscription churn analysis, but its implemented behavior materially falls short of the declared primary capability. Each criterion lookup uses crit.get("column", ""), while none of the criteria defines a column, so ordinary inputs produce all-zero scores and effectively retain input order. The output is only a flat ranked list; no churn-reason or renewal-date grouping is implemented. Local CSV reading and report writing are supporting implementation details, and there is no evidence of unrelated network access or external actions.

Content

No source excerpt is available for this finding.

Skill declares no tool scope ('permissions' or 'allowed-tools') but code capabilities were detected: file_write.

Medium
Category
MCP Least Privilege
Confidence
70% confidence
Finding

Without declared permissions the skill's intent is opaque and cannot be validated.

Content

No source excerpt is available for this finding.

The claimed churn-risk analysis produces zero scores for ordinary account CSVs instead of evaluating the advertised risk criteria.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

Every criterion lacks a column mapping, so the scoring loop reads the empty-string column name and defaults missing values to zero; ordinary account CSVs therefore receive identical zero scores regardless of churn risk. This is an analytical integrity failure in the skill's core function: customer-success teams could mistake these outputs for valid risk assessments and misprioritize retention work, although there is no evidence of deliberate manipulation.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.