ClawHub
Back to skill

Security audit

Phy Influencer Research

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed influencer-research workflow that searches public web sources and creates an Excel report, with some scope and file-output assumptions users should review.

Before installing, be aware that this skill may run broad public web searches and create an Excel file automatically. Provide explicit niche, region, platforms, target count, and output expectations when accuracy matters, and avoid autonomous mode if you want to approve scope before research begins.

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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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 (4)

Vague Triggers

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The trigger language is broad enough to activate on many generic requests involving lists, databases, or outreach-related research, which increases the chance the skill runs in contexts the user did not explicitly intend. Because the skill performs large-scale web research and file generation, unintended invocation could cause unnecessary data collection, misleading outputs, or downstream actions based on incorrect assumptions.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The skill instructs automatic Excel creation under a constructed output path but does not mention confirmation, overwrite handling, or safe filename sanitization. In agent environments with filesystem access, this can lead to silent file creation, accidental overwrites, or writing files with unexpected names derived from user-controlled inputs like region and niche.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

Inferring region from language without opt-in can produce materially incorrect targeting assumptions and may encode biased defaults, especially for multilingual users or global campaigns. Since the skill compiles outreach-oriented influencer datasets, these silent assumptions can distort results and cause the agent to act on inaccurate geographic scope.

Content

No source excerpt is available for this finding.

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
95% confidence
Finding

The instruction to proceed without confirmation enables the agent to make substantive decisions about scope, platforms, volume, and geography autonomously. In a skill that performs external research and creates output files, this increases the risk of over-collection, wasted resources, and outputs that do not match user intent.

Content

Scanner excerpt · SKILL.md (reported line 180)May include surrounding context.

md
- Default region based on language (English → North America, Chinese → China/SEA)
- Default to all 5 platforms
- Default to 30-50 per platform target
- Proceed without confirmation prompts

---

Static analysis

No suspicious patterns detected.