ClawHub
Back to skill

Security audit

agent-analytics-autoresearch

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed analytics-assisted growth workflow that writes local review artifacts and only changes live experiments after explicit user approval.

Install this only if you are comfortable letting the agent read or collect analytics for the named project and save those snapshots locally. Use least-privilege Agent Analytics access, review snapshot files before sharing or committing them, and only approve live experiment wiring when you intend the agent to modify product or experiment setup.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Tp4

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The declared description promises a full growth-research and experimentation workflow, including preserving product truth, generating and evaluating candidate variants, ranking them, and outputting two A/B test variants. The supplied code does none of that. It only gathers analytics data from the Agent Analytics CLI and writes the results to text files. While snapshot collection is one supporting part of the declared functionality, the code chunk’s actual behavior is much narrower and does not implement the primary capabilities described. Therefore the description materially overstates what this code does.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
96% confidence
Finding

The skill explicitly requires and invokes npx to download and run a package from the registry, and the metadata only states anyBins: npx rather than pinning the execution environment itself. Even though the CLI package version is pinned later, relying on npx as a network-fetching execution path increases supply-chain risk because execution depends on the caller's npm/npx toolchain and registry trust at runtime.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The template directs users to collect analytics snapshots, event logs, and experiment data and save them under a local data directory, but it provides no safeguards around sensitive contents, retention, access controls, or redaction. In the context of an agent skill that may be run automatically, this increases the chance of storing user behavior data, project identifiers, or command error output in plaintext where it could be exposed, mishandled, or later exfiltrated.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.