Back to skill

Security audit

nacos-skill-registry

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed Nacos registry helper that can install, update, and publish skills, with some credential and backup handling users should review before use.

Before installing, make sure you intend this skill to let an agent use nacos-cli against your configured Nacos registry. Confirm the active profile, namespace, and auth mode before any upload, review, or release, and avoid leaving private skill backups in /tmp on shared machines.

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
  • Rogue AgentSelf-Modification, Session Persistence
  • 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
Findings (5)

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 506)May include surrounding context.

md
A valid skill directory must contain a `SKILL.md` file with proper frontmatter (name, description). Confirm the path:

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The activation guidance includes generic phrases like "how do I do X" and broad requests such as "what skills are available," which are common in everyday conversation and could cause the skill to be invoked too often. The document does not provide exclusion conditions or negative examples to narrow when this skill should activate versus when general assistance should be used.

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
79% confidence
Finding

The skill encourages reading and relying on persistent local environment variables and profile files for authentication and connection state. This creates session persistence and ambient-authority behavior: an agent may act using preexisting credentials or configuration without fresh user confirmation, potentially against the wrong Nacos environment or namespace. In a registry-management skill, that can lead to unintended access, publishing, or data disclosure.

Content

Scanner excerpt · SKILL.md (reported line 102)May include surrounding context.

nacos-cli uses a profile mechanism. Profile configs are stored in ~/.nacos-cli/<profile>.conf. The default profile is default (i.e., ~/.nacos-cli/default.conf).

Before asking the user to create or edit a profile, check whether the environment already provides a complete sts-hiclaw setup:

bash

Session Persistence

Medium
Category
Rogue Agent
Confidence
71% confidence
Finding

This workflow persists local copies and backups of skills under predictable filesystem locations such as ~/.skills and /tmp. While operationally useful, it increases retention of potentially sensitive skill contents and local modifications, and /tmp backups may be accessible to other local users or remain after the task ends. In the context of team or private skill registries, that can expose unpublished or internal skill logic.

Content

Scanner excerpt · SKILL.md (reported line 402)May include surrounding context.

Option A: Discard local changes, use remote version

This is the simpler path. Back up the local version first, then overwrite with the remote:

bash
# 1. Back up the current local skill

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · SKILL.md (reported line 492)May include surrounding context.

B) Merge local and remote changes, then publish as a new version

text

**Never overwrite local changes without asking.** Always back up before any destructive operation.

## How to Help Users Publish Skills to Nacos