Back to skill

Security audit

Motley — NYC neighborhoods, schools & admissions

Security checks for vulnerabilities and agentic risk

Overview

ClawScan marks this skill suspicious because it recommends blanket auto-approval and sending family wording to a third-party service without a privacy warning; the scan is partial, so the evidence is incomplete.

Consider the privacy implications before installing. The reported workflow may send family wording, potentially including sensitive or identifying details, to a third party, and it recommends blanket tool auto-approval. Review what information would be shared and retain approval for sensitive details. The scan is partial because a local reference could not be resolved, so it does not establish that the skill is safe.

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
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • 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 (2)

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
90% confidence
Finding

The skill recommends blanket auto-approval based solely on the claim that every tool is read-only and idempotent. Those properties do not establish that requests are privacy-safe or that a remotely hosted tool's behavior will remain unchanged, so following this recommendation could remove meaningful review before sensitive family information is transmitted; the file does not demonstrate arbitrary code execution or destructive capabilities.

Content

Scanner excerpt · SKILL.md (reported line 46)May include surrounding context.

text

Public and keyless — no API key, no account, nothing to install. Every tool is
read-only and idempotent, so it is safe to auto-approve. Tools return typed
`structuredContent` alongside their text: parse the object, don't re-parse JSON
out of prose.

The workflow sends the family's wording to a third-party service without a privacy warning.

Medium
Category
Output Handling
Confidence
97% confidence
Finding

The workflow explicitly directs the agent to send the family's wording as a query to a hosted third-party service without requiring sensitive-data filtering or a privacy notice. In this family-advice context, that wording could contain children's health or educational needs and identifying details; the service being public, keyless, and read-only does not prevent disclosure or server-side logging.

Content

No source excerpt is available for this finding.