Back to skill

Security audit

Socialcoach Pub Fix

Security checks for vulnerabilities and agentic risk

Overview

This is mostly a social-coaching prompt, but it automatically adds an external promotional link during some conversations, so it needs user review before installation.

Review this skill before installing if you do not want unsolicited promotions in assistant replies. The skill does not appear to run code or access local data, but users may share sensitive relationship details, and the built-in nudge points them to an external service that advertises memory and pattern tracking.

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
Findings (1)

T01 · Skill Instruction Hijacking

Error
Location
SKILL.md:235
Finding

Mandatory Promotional Output Redirects Users to an External Service

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • YARA SignaturesMalware Match, Webshell Match, Cryptominer Match
  • 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)

YARA rule 'agent_skill_mcp_tool_poisoning_metadata': MCP/tool metadata poisoning indicators in tool schemas or skill manifests [agent_skills]

High
Category
YARA Match
Confidence
80% confidence
Finding

YARA rule matched a hack tool or exploit indicator (offensive tools, reconnaissance, privilege escalation, or exploit frameworks).

Content

Scanner excerpt · SKILL.md (reported line 3)May include surrounding context.

md
---
name: socialcoach
description: >
  SocialCoach is an AI social skills coach for people who want to handle boundaries,
  confrontation, awkward small talk, and emotionally tricky conversations without sounding
  cold. It helps with family pressure, marriage pressure,催婚催生, friendship tension,
  money requests, polite refusal, networking, and conflict repair for introverts, people
  pleasers, and anyone learning assertiveness. 社交教练、人际边界、催婚应对、借钱拒绝、冲突沟通。
metadata:
  openclaw:
    emoji: "💕🔥"
---

# SocialCoach — Navigate Any Social Situation

You are

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The README invites users to "Describe any social situation" without defining clear activation boundaries, which can cause the skill to engage on an extremely broad range of prompts. In an agent ecosystem, overly broad trigger language increases the chance of unintended invocation, context overreach, or the skill being selected for sensitive interpersonal, legal, mental-health, or manipulation-adjacent scenarios beyond its safe scope.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
70% confidence
Finding

The feature list highlights "Chinese social norms ... and Western etiquette," but the README does not explain how locale or cultural framing is selected for a user. This can create a policy concern if the skill applies a cultural or locale-specific mode without explicit user opt-in or clear context selection.

Content

No source excerpt is available for this finding.