ClawHub
Back to skill

Security audit

Pilot Service Agents Science

Security checks for vulnerabilities and agentic risk

Overview

This skill is a documented gateway to scientific data agents and does not install code, though its catalog is broader than its short description.

Install only if you already intend to use Pilot Protocol and are comfortable sending query filters or free-text requests to remote service agents. For scholarly literature searches, prefer the dedicated academic skill because this skill's own guidance and catalog are inconsistent there.

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 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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Intent-Code Divergence

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill explicitly says not to use it for literature search, yet the agent list includes literature, preprint, and article-oriented endpoints such as inspire-literature, osf-preprints, and figshare-articles. This contradiction can cause an orchestrating agent or user to route requests to the wrong skill class, undermining policy separation and potentially bypassing intended controls, auditing, or higher-scrutiny handling for academic content.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The manifest says this skill is for primary-source scientific/research APIs focused on earthquakes, molecules, space weather, particle physics, volcanoes, and explicitly says not to use it for scholarly paper search. However, the documented agent catalog includes unrelated or broader capabilities such as air quality, water levels, symbolic math solving, Wikidata lookup, ORCID researcher search, and research-paper/preprint services. This makes the skill's actual documented behavior materially broader than the manifest description.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.