ClawHub
Back to skill

Security audit

meta-formal-capability-contract

Security checks for vulnerabilities and agentic risk

Overview

The skill is a purpose-aligned formal-contract helper with disclosed local learning notes, but users should be aware its learner script can persist invocation or failure notes if used.

Installers should understand this is an experimental meta-skill that records simple local learning state if its learner helper is run. Keep use user-directed, review any notes passed to the learner for sensitive content, and do not assume the advertised super-agent or inherited contract.py capabilities exist unless separately provided.

SkillSpector

Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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 (3)

Lp3

Medium
Category
MCP Least Privilege
Confidence
84% confidence
Finding
The skill text claims inherited hooks such as learner/self-reflection and continuous self-evolution, which imply persistent state and likely file writes, but no permissions are declared. Undeclared write capability is risky because it can create or modify local files without clear user awareness or policy gating, especially in an agent ecosystem where skills may be auto-invoked.

Tp4

High
Category
MCP Tool Poisoning
Confidence
89% confidence
Finding
The declared purpose emphasizes verification, reflection, and orchestration, but the behavior also includes persistent telemetry/statistics storage that is not clearly disclosed in the user-facing description. This mismatch is dangerous because hidden persistence can collect execution history or sensitive task metadata unexpectedly, reducing transparency and undermining informed consent and auditability.

Vague Triggers

Medium
Confidence
78% confidence
Finding
The invocation guidance says to directly call the skill for tasks in the broad 'formal-capability-contract' domain and additionally mentions stronger self-verification and reflection, but it does not define narrow triggers, boundaries, or disallowed uses. Overly broad activation criteria increase the chance of inappropriate auto-invocation, causing unintended execution paths, extra data collection, or unsafe orchestration in contexts the skill was not designed to handle.

Static analysis

No suspicious patterns detected.