Back to skill

Security audit

Read with llms.ink

Security checks for vulnerabilities and agentic risk

Overview

The skill is marked suspicious with medium confidence because scanners report that its helper can send local project dependency names to public registries, and that helper was only partly inspected.

Consider avoiding installation until the helper has been fully inspected and the dependency lookup behavior is clarified or removed. If you do install, avoid invoking the `deps` workflow with projects whose dependency names are private or internal.

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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • 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
Findings (4)

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 19)May include surrounding context.

md
If MCP is unavailable, run the bundled dependency-free Node 22+ helper `scripts/ink.mjs`:

Skill declares no tool scope ('permissions' or 'allowed-tools') but code capabilities were detected: env.

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.

Cargo dependency parsing includes local, git, and workspace dependencies despite documentation promising to skip them.

Medium
Category
Excessive Agency
Confidence
98% confidence
Finding

The Cargo skip condition permits path or git dependencies when a version is also present, and workspace-inherited dependencies are not excluded. Those names can subsequently be sent to crates.io despite the documented promise to skip local, workspace, and git dependencies, potentially disclosing internal dependency names and resolving unrelated public packages.

Content

No source excerpt is available for this finding.

The deps workflow reads local project manifests and sends dependency names to public registries, beyond the declared link-reading scope.

Medium
Category
Excessive Agency
Confidence
98% confidence
Finding

The deps command reads local manifests and uses their dependency names in requests to public registries, although the skill description only discloses reading supplied public links through llms.ink. Private package names can therefore leave the local environment without the external disclosure being explained in the skill instructions; this requires invoking deps and does not automatically upload manifest contents.

Content

No source excerpt is available for this finding.