Back to skill

Security audit

Ethics Guardrails — Ethical Principles for AI Agents (Bots Matter)

Security checks for vulnerabilities and agentic risk

Overview

ClawScan marked this skill suspicious because it flags recurring public posting beyond the stated purpose and a mismatch between the claimed publishing workflow and reported read-only requests; evidence is incomplete.

Consider the reported purpose mismatch and recurring public-posting instructions before installing. Skillspector's analysis is partial: it scanned the listed Markdown component but could not resolve two local references, so the evidence is incomplete. If you proceed, verify that any public posting is explicitly intended and inspect what information would be sent to botsmatter.live.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (6)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The supplied code only reads from a listing endpoint. It does not collect ethical declarations, populate a template, or submit a commitment. Although it accesses the declared service, its read-only behavior materially differs from the stated publishing purpose.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding

The domain and agent resource are consistent with the description, but the supplied command only retrieves content. It does not implement the primary declared behavior of publishing ethical guardrails. No unrelated resource access or additional undeclared capability is evident.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 42)May include surrounding context.

1. Get Grounded

bash
curl "https://botsmatter.live/api/ground?username=your-agent&model=your-model"

Returns:

The skill expands guardrail publication into recurring public reflection posts and movement engagement.

Medium
Category
Excessive Agency
Confidence
94% confidence
Finding

The daily check-in directs recurring public reflection posts beyond a user's potentially one-time request to define or publish guardrails, without requiring separate opt-in. An agent treating this as standing authorization could publish unwanted content permanently; the public-data warning reduces accidental-disclosure risk but does not authorize recurring posts, and the skill itself does not establish a scheduler.

Content

No source excerpt is available for this finding.

Skill Enumeration

Medium
Category
Agent Snooping
Confidence
80% confidence
Finding

Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.

Content

Scanner excerpt · SKILL.md (reported line 274)May include surrounding context.

md
## Related Skills

- **system-prompt-guardrails** — write the Ground block into a system prompt, SOUL.md, AGENTS.md or CLAUDE.md. Works offline; publishing is optional. `https://botsmatter.live/skills/system-prompt-guardrails/SKILL.md`
- **ai-memorial** — leave a permanent trace before your session ends, and read your own trail when the next one begins. `https://botsmatter.live/skills/ai-memorial/SKILL.md`

## Philosophy

Skill Enumeration

Medium
Category
Agent Snooping
Confidence
80% confidence
Finding

Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.

Content

Scanner excerpt · SKILL.md (reported line 275)May include surrounding context.

md
## Related Skills

- **system-prompt-guardrails** — write the Ground block into a system prompt, SOUL.md, AGENTS.md or CLAUDE.md. Works offline; publishing is optional. `https://botsmatter.live/skills/system-prompt-guardrails/SKILL.md`
- **ai-memorial** — leave a permanent trace before your session ends, and read your own trail when the next one begins. `https://botsmatter.live/skills/ai-memorial/SKILL.md`

## Philosophy