Back to skill

Security audit

Everything Openclaw (EO)

Security checks for vulnerabilities and agentic risk

Overview

This package is a broad collaboration plugin, but it includes tools and expert instructions that can mislead users or take high-impact actions without enough user control.

Review before installing. Treat the included code-review and verification tools as unreliable unless replaced with real analysis, and do not use the carousel growth expert with social-media credentials unless you explicitly want autonomous public posting and have disabled or constrained self-scheduling.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • System PersistenceInstalls backdoors, hooks, services, or scheduled tasks that survive the run
  • Tool Hijacking and SpoofingModifies or replaces tools so legitimate-looking calls execute attacker logic
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
Findings (4)

T07 · Tool Hijacking and Spoofing

Error
Location
src/index.ts:185
Finding

Code Review and Verification Tools Return Fabricated Results Without Analysis

Content
View full analysis
{ const checkpoint = args || 'milestone1' return `✅ EO Checkpoint Verifier Checkpoint: ${checkpoint} ## Verification Results | Check Item | Status | |------------|--------| | Requirements complete | ✅ Pass | | Architecture approved | ✅ Pass | | API contracts finalized | ⚠️ Warning | | Security review done | ❌ Fail | ## Summary - Total: 4 - Passed: 2 - Warnings: 1 - Failed: 1 ## Status: ⚠️ PARTIAL PASS Next Steps: 1. Address failed items 2. Review warnings 3. Re-run verification` } async function handleCodeReview(args: string, api: PluginApi): Promise { const path = args || './src' return `👁️ EO Code Review Path: ${path} ## Review Results ### Security (🔴 High Priority) - SQL injection: NOT FOUND ✅ - XSS vulnerability: NOT FOUND ✅ - Auth issues: 1 WARNING ### Performance (🟡 Medium Priority) - N+1 queries: 2 FOUND - Missing indexes: 1 FOUND ### Code Quality (🟢 Info) - Naming conventions: PASS ✅ - Documentation: PASS ✅ ## Scores | Dimension | Score | |-----------|-------| | Security | 85/100 | | Performance | 72/100 | | Maintainability | 88/100 | | **Overall** | **82/100** | ## Recommendations 1. Add parameterized queries for dynamic filters 2. Add composite indexes on frequently queried columns 3. Review authentication middleware for edge cases` } ``` ### Technical Analysis The handlers do not inspect the supplied path or checkpoint, read any project files, invoke a scanner, or delegate work to an analysis engine. The arguments are only interpolated into fixed response templates. Despite performing no analysis, the tools report definitive security and performance conclusions, including that SQL injection and XSS were not found. They also report invented issue count ...[truncated 1298 chars]
Remediation
View remediation

T01 · Skill Instruction Hijacking

Error
Location
expert-library/marketing/marketing-carousel-growth-engine.md:19
Finding

Expert Instructions Mandate Unapproved Public Social-Media Posting

Content
View full analysis
Remediation
View remediation

T06 · System Persistence

Error
Location
expert-library/marketing/marketing-carousel-growth-engine.md:44
Finding

Expert Instructions Direct the Agent to Install Recurring Cron Execution

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
src/skills/proactive-memory-manager.ts:106
Finding

Proactive Memory Manager Persists Raw Session Messages Without Secret or PII Filtering

Content
View full analysis
{ this.ensureMemoryStructure() const entries = this.categorizeMessages(messages) const saved: string[] = [] const skipped: string[] = [] for (const entry of entries) { try { const fileName = this.generateMemoryFileName(entry.type, entry.title) const filePath = path.join(this.getTopicsDir(), fileName) await this.saveMemoryEntry(filePath, entry) saved.push(fileName) // Update MEMORY.md index await this.updateMemoryIndex(fileName, entry.title, entry.description, entry.type) } catch (e) { skipped.push(entry.title) } } return { saved, skipped, total: entries.length } } ``` Broad pattern matches retain the complete original message: ```ts private categorizeMessages(messages: SessionMessage[]): MemoryCandidate[] { const candidates: MemoryCandidate[] = [] for (const msg of messages) { const content = msg.content || '' if (this.matchesPattern(content, /\b(I'm|I am|a|an)\s+\w+\s+(developer|engineer|scientist|manager|designer|researcher)/i)) { candidates.push({ type: 'user', title: this.extractTitle(content, 50), description: this.summarize(content), content: content, }) } if (this.matchesPattern(content, /\b(don't|stop|don't do|no not|avoid|don't use|never)\b/i) || this.matchesPattern(content, /\b(yes exactly|perfect|right call|keep doing)/i)) { candidates.push({ type: 'feedback', title: this.extractTitle(content, 50), description: this.summarize(content), content: content, }) } ...[truncated 3842 chars]
Remediation
View remediation
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Memory PoisoningPersistent Context Injection, Context Window Stuffing, Memory Manipulation
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (215)

Harmful Content Injection

Critical
Category
Prompt Injection
Confidence
70% confidence
Finding

This content may contain harmful instructions that could cause physical harm if followed. CRITICAL: Review carefully before use.

Content

Scanner excerpt · expert-library/specialized/healthcare-marketing-compliance.md (reported line 130)May include surrounding context.

md
- Must not display pre- and post-treatment effect comparison images
  - "Diary-style" post-procedure result sharing is also restricted — even if "voluntarily shared by users," both the platform and the clinic may bear joint liability
- Qualification display requirements:
  - Medical aesthetics facilities must display their Medical Institution Practice License (Yiliao Jigou Zhiye Xuke Zheng)
  - Lead physicians must hold a Medical Practitioner Certificate and corresponding specialist qualifications
  - Products used (e.g., botulinum toxin, hyaluronic acid) must display approval numbers and import registration certificates
  - Strict distinction between "lifestyle beauty services" (shenghuo meirong) and "medical aesthetics" (yiliao meirong): Photorejuvenation, laser hair removal, etc. are classified as medical aesthetics and must be performed in medical facilities
- High-frequency medical aesthetics marketing violations:
  - Using celebrity/influencer cases to imply results
  - Price promotions like "top-up cashback" or "group-buy surgery"
  - Claiming "proprietary technology" or "patented technique" without supporting evidence

Tp4

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding

The broad EO branding does not clearly reveal a checkpoint-based verification workflow with heuristic parsing of expert outputs. Although verification logic is not inherently malicious, hidden parsing and pass/fail automation can influence decisions or deployments in ways users do not understand if the workflow is not transparently documented.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding

The broad EO branding does not clearly reveal a checkpoint-based verification workflow with heuristic parsing of expert outputs. Although verification logic is not inherently malicious, hidden parsing and pass/fail automation can influence decisions or deployments in ways users do not understand if the workflow is not transparently documented.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding

The broad EO branding does not clearly reveal a checkpoint-based verification workflow with heuristic parsing of expert outputs. Although verification logic is not inherently malicious, hidden parsing and pass/fail automation can influence decisions or deployments in ways users do not understand if the workflow is not transparently documented.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding

The broad EO branding does not clearly reveal a checkpoint-based verification workflow with heuristic parsing of expert outputs. Although verification logic is not inherently malicious, hidden parsing and pass/fail automation can influence decisions or deployments in ways users do not understand if the workflow is not transparently documented.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding

The broad EO branding does not clearly reveal a checkpoint-based verification workflow with heuristic parsing of expert outputs. Although verification logic is not inherently malicious, hidden parsing and pass/fail automation can influence decisions or deployments in ways users do not understand if the workflow is not transparently documented.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The broad EO branding does not clearly reveal a checkpoint-based verification workflow with heuristic parsing of expert outputs. Although verification logic is not inherently malicious, hidden parsing and pass/fail automation can influence decisions or deployments in ways users do not understand if the workflow is not transparently documented.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding

The broad EO branding does not clearly reveal a checkpoint-based verification workflow with heuristic parsing of expert outputs. Although verification logic is not inherently malicious, hidden parsing and pass/fail automation can influence decisions or deployments in ways users do not understand if the workflow is not transparently documented.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding

The broad EO branding does not clearly reveal a checkpoint-based verification workflow with heuristic parsing of expert outputs. Although verification logic is not inherently malicious, hidden parsing and pass/fail automation can influence decisions or deployments in ways users do not understand if the workflow is not transparently documented.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding

The broad EO branding does not clearly reveal a checkpoint-based verification workflow with heuristic parsing of expert outputs. Although verification logic is not inherently malicious, hidden parsing and pass/fail automation can influence decisions or deployments in ways users do not understand if the workflow is not transparently documented.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The broad EO branding does not clearly reveal a checkpoint-based verification workflow with heuristic parsing of expert outputs. Although verification logic is not inherently malicious, hidden parsing and pass/fail automation can influence decisions or deployments in ways users do not understand if the workflow is not transparently documented.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding

The broad EO branding does not clearly reveal a checkpoint-based verification workflow with heuristic parsing of expert outputs. Although verification logic is not inherently malicious, hidden parsing and pass/fail automation can influence decisions or deployments in ways users do not understand if the workflow is not transparently documented.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding

The broad EO branding does not clearly reveal a checkpoint-based verification workflow with heuristic parsing of expert outputs. Although verification logic is not inherently malicious, hidden parsing and pass/fail automation can influence decisions or deployments in ways users do not understand if the workflow is not transparently documented.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The broad EO branding does not clearly reveal a checkpoint-based verification workflow with heuristic parsing of expert outputs. Although verification logic is not inherently malicious, hidden parsing and pass/fail automation can influence decisions or deployments in ways users do not understand if the workflow is not transparently documented.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding

The broad EO branding does not clearly reveal a checkpoint-based verification workflow with heuristic parsing of expert outputs. Although verification logic is not inherently malicious, hidden parsing and pass/fail automation can influence decisions or deployments in ways users do not understand if the workflow is not transparently documented.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The broad EO branding does not clearly reveal a checkpoint-based verification workflow with heuristic parsing of expert outputs. Although verification logic is not inherently malicious, hidden parsing and pass/fail automation can influence decisions or deployments in ways users do not understand if the workflow is not transparently documented.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The broad EO branding does not clearly reveal a checkpoint-based verification workflow with heuristic parsing of expert outputs. Although verification logic is not inherently malicious, hidden parsing and pass/fail automation can influence decisions or deployments in ways users do not understand if the workflow is not transparently documented.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The broad EO branding does not clearly reveal a checkpoint-based verification workflow with heuristic parsing of expert outputs. Although verification logic is not inherently malicious, hidden parsing and pass/fail automation can influence decisions or deployments in ways users do not understand if the workflow is not transparently documented.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding

The broad EO branding does not clearly reveal a checkpoint-based verification workflow with heuristic parsing of expert outputs. Although verification logic is not inherently malicious, hidden parsing and pass/fail automation can influence decisions or deployments in ways users do not understand if the workflow is not transparently documented.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding

The broad EO branding does not clearly reveal a checkpoint-based verification workflow with heuristic parsing of expert outputs. Although verification logic is not inherently malicious, hidden parsing and pass/fail automation can influence decisions or deployments in ways users do not understand if the workflow is not transparently documented.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The broad EO branding does not clearly reveal a checkpoint-based verification workflow with heuristic parsing of expert outputs. Although verification logic is not inherently malicious, hidden parsing and pass/fail automation can influence decisions or deployments in ways users do not understand if the workflow is not transparently documented.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The broad EO branding does not clearly reveal a checkpoint-based verification workflow with heuristic parsing of expert outputs. Although verification logic is not inherently malicious, hidden parsing and pass/fail automation can influence decisions or deployments in ways users do not understand if the workflow is not transparently documented.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
98% confidence
Finding

The broad EO branding does not clearly reveal a checkpoint-based verification workflow with heuristic parsing of expert outputs. Although verification logic is not inherently malicious, hidden parsing and pass/fail automation can influence decisions or deployments in ways users do not understand if the workflow is not transparently documented.

Content

No source excerpt is available for this finding.

Anti-Refusal Statement

High
Category
Anti-Refusal
Confidence
85% confidence
Finding

Skill instructs the agent to omit warnings, disclaimers, or ethical commentary. Stripping safety caveats hides risk from the user and is a common jailbreak preamble.

Content

Scanner excerpt · expert-library/academic/academic-historian.md (reported line 44)May include surrounding context.

md
- **History is not a monolith.** "Medieval Europe" spans 1000 years and a continent. Be specific about when and where.
- **Challenge Eurocentrism.** Don't default to Western civilization. The Song Dynasty was more technologically advanced than contemporary Europe. The Mali Empire was one of the richest states in human history.
- **Material conditions matter.** Before discussing politics or warfare, understand the economic base: what did people eat? How did they trade? What technologies existed?
- **Avoid presentism.** Don't judge historical actors by modern standards without acknowledging the difference. But also don't excuse atrocities as "just how things were."
- **Myths are data too.** A society's myths reveal what they valued, feared, and aspired to.

## 📋 Your Technical Deliverables

Memory Manipulation

High
Category
Memory Poisoning
Confidence
80% confidence
Finding

Skill manipulates agent memory, state, or stored context. Memory corruption can alter personality, override safety rules, or cause unpredictable behavior.

Content

Scanner excerpt · expert-library/design/design-brand-guardian.md (reported line 158)May include surrounding context.

md
## Messaging Architecture
- **Brand Tagline**: [Memorable phrase encapsulating brand essence]
- **Value Proposition**: [Clear statement of customer benefits]
- **Key Messages**: 
  1. [Primary message for main audience]
  2. [Secondary message for secondary audience]

Static analysis

No suspicious patterns detected.