Back to skill

Security audit

capability-match

Security checks for vulnerabilities and agentic risk

Overview

This skill locally reads installed skill descriptions to recommend a matching skill, with no hidden persistence, network transfer, or destructive behavior found.

Install only if you are comfortable with the skill reading the SKILL.md files for all installed OpenClaw skills and showing skill names, descriptions, and trigger examples in recommendations. Treat its suggestions as advisory and choose the next skill yourself.

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
  • Rogue AgentSelf-Modification, Session Persistence
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (6)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The trigger examples are very broad natural-language phrases such as "what skill should i use for" and "use capability-match to," which can overlap with ordinary conversation and cause unintended activation. In a router skill that scans installed skills and influences downstream tool selection, accidental invocation increases the chance of misrouting user intent or exposing skill metadata more often than expected.

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
60% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · README.md (reported line 24)May include surrounding context.

Example

You: capability-match: create a dashboard from my CSV data

Response:

text

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
70% confidence
Finding

Without declared permissions the skill's intent is opaque and cannot be validated.

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
60% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · SKILL.md (reported line 17)May include surrounding context.

Simply describe what you want to do:

text
Use capability match to: "create a PowerPoint presentation from sales data"

Or directly:

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
84% confidence
Finding

The trigger extraction logic recognizes only hard-coded English phrases such as "orchestrate" and "what skill should i use for". This creates a language policy issue because users of other languages are implicitly unsupported without any documented opt-in, fallback, or locale selection.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The description advertises an AI-powered router that 'analyzes your request and recommends the best installed skill' without defining clear boundaries, allowed domains, or trigger constraints. Broad routing language increases the chance that the skill is invoked for unintended requests or becomes a confused-deputy component that forwards sensitive or risky tasks to other installed skills.

Content

No source excerpt is available for this finding.