Security audit

Ai Trend Curation

Security checks for vulnerabilities and agentic risk

Overview

The skill is a disclosed AI-news curation workflow that posts curated public X content to Slack, with no embedded executable payload or hidden persistence, but users should confirm Slack destination and note the missing referenced script.

Install only if you want an agent to curate public AI tweets and post the result to Slack. Before use, confirm which Slack workspace/channel openclaw message send will target, review the generated summary before sending, and verify where the missing scripts/ai_trends.js implementation is supposed to come from.

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
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 (4)

Description-Behavior Mismatch

Medium

Category: Not specified by scanner
Confidence: 98% confidence
Finding: The manifest claims the skill will "Curate and post AI trend tweets from X (Twitter) with quote suggestions," which implies publishing on X. However, the workflow documentation says the skill collects tweets from X and then posts the curated output to Slack via openclaw message send, making Slack posting the actual output behavior. This is a semantic mismatch in the skill's declared purpose versus documented behavior.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 91% confidence
Finding: The document is written in Japanese and later mandates a specific Japanese persona and style for quote suggestions, but it does not present this as an opt-in or explain a justified region-specific limitation. This creates a language/locale policy concern because the skill appears to force a specific language behavior by default.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 89% confidence
Finding: The skill instructs the agent to send curated content to Slack via an external messaging command without any user-facing warning, approval checkpoint, or data-sharing notice. That can cause unintended disclosure of generated summaries, tweet selections, or internal context to a third-party service, especially if the skill is run in environments where outbound posting should be explicit and consented.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: These lines prescribe output in a specific Japanese persona (e.g., first-person '僕') and style, which is a natural-language policy issue when no user opt-in or alternative locale choice is offered. The requirement appears mandatory rather than contextual or configurable.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.