Back to skill

Security audit

skill-architecture-standard

Security checks for vulnerabilities and agentic risk

Overview

This is a disclosed Indonesian-language architecture reference for building and reviewing OpenClaw skills, with no executable code or hidden data access.

Install only if you want an Indonesian reference standard that influences how agents create or review OpenClaw skills. Treat its tool, memory, and enforcement sections as design guidance for other skills, and continue requiring normal user approval for high-risk actions such as installs, config edits, deletions, credentials, or external calls.

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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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
Findings (4)

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The manifest presents this as a passive reference, but the body defines agentic execution behavior including acting, verifying, recovering, and selecting tools. That mismatch can cause an orchestrator or reviewer to load a seemingly harmless reference skill that in practice steers operational behavior beyond its declared purpose, increasing the chance of unintended actions.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The manifest description is written as a directive in Indonesian ('Gunakan sebagai referensi...') and the rest of the document consistently prescribes behavior in Indonesian, with no indication that users may choose another language. This creates a locale/language policy concern because the skill appears to require a specific language without user opt-in or justification as a region-specific skill.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The document states that it does not perform domain tasks, but later instructs the agent to execute actions and choose tools directly. This contradiction is dangerous because it weakens trust boundaries: operators may treat the skill as non-operational while it still influences execution paths, verification logic, and system interaction.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The skill includes guidance to use terminal, web, and file-editing capabilities even though its stated role is only an architecture reference. In a skill-loading system, broad operational instructions inside a meta-skill can normalize or indirectly authorize risky actions in contexts where no domain-specific justification or safety boundary exists.

Content

No source excerpt is available for this finding.