ClawHub
Back to skill

Security audit

Federation Ethical Hand

Security checks for vulnerabilities and agentic risk

Overview

The skill is not malicious, but it asks agents to persist unspecified ethics artifacts and metrics under the user's home directory without clear limits or user control.

Install only if you are comfortable with an ethics helper that may change how broad categories of requests are handled and may write local artifacts or metrics. Before use, confirm what is saved under ~/.openclaw/hands/federation_ethic_hand, whether prompt content is included, and how to disable or delete that data.

A.I.G

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

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Rogue AgentSelf-Modification, Session Persistence
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill metadata and description frame activation around broad ethical themes without a precise gating rule, which can cause the skill to activate for a wide range of normal requests. Over-broad activation is dangerous because it can unexpectedly alter model behavior, inject extra policy layers, or influence responses outside the user's intent, especially when the skill also performs persistence and logging.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The trigger list includes a catch-all condition, 'Any ethically ambiguous decision with multiple legitimate viewpoints,' which is highly subjective and effectively permits near-unbounded invocation. This creates a control-plane risk because the skill can be loaded in many contexts where users did not request it, changing reasoning flow, web-fetch behavior, and output framing unexpectedly.

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
96% confidence
Finding

Phase 5 explicitly instructs the agent to persist artefacts and log metrics, and the listed key files indicate storage under a user home directory. Session persistence is dangerous because sensitive request content, ethical classifications, or derived metadata can accumulate across sessions, enabling unintended retention, privacy leakage, or later misuse if the files are accessed by other tools or users.

Content

Scanner excerpt · SKILL.md (reported line 34)May include surrounding context.

md
2. **Phase 2 – Execute**: Apply cultural sensitivity level to all web searches and fetches
3. **Phase 3 – Deliberate**: Run the 3 Ethics Committee questions; tag all sources with cultural lens
4. **Phase 4 – Synthesise**: Generate multi-perspective viewpoints (1/3/5 based on `perspective_count` setting); label each lens; flag conflicts
5. **Phase 5 – Output & Persist**: Write artefacts; log ethics metrics

## Key Files

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill directs the agent to 'Write artefacts; log ethics metrics' and references persistent files, but it does not clearly warn the user that data or metadata may be written to storage. Hidden persistence is risky because it can retain sensitive prompts, derived judgments, or usage metadata beyond the session, violating user expectations and potentially organizational data-handling requirements.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.