X Search (Grok API)

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:32 PM UTC
• Version:
  1.0.0

Overview

This is a disclosed X/Twitter search helper; users should be aware it can save or overwrite local Markdown files and optionally send scheduled results to messaging platforms.

Install only if you are comfortable sending X/Twitter queries and returned content to xAI through your API key. Use a dedicated output directory because --output can overwrite same-day files, and enable cron or Telegram/Slack/Discord delivery only for searches whose results you are comfortable storing or sharing in those services.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Rogue Agent: Self-Modification, Session Persistence
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection

Findings (6)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 96% confidence
• Finding: The skill advertises capabilities that include environment access, network use, shell execution, and file writing, but does not declare any permissions or constraints in the manifest. This creates a transparency and governance gap: an agent or reviewer may treat the skill as a simple search formatter while it can access secrets, make outbound requests, and write persistent files.

Description-Behavior Mismatch

Medium

• Confidence: 92% confidence
• Finding: The manifest frames the skill as a search tool that outputs Markdown, but the documentation also allows saving to arbitrary user-supplied paths and overwriting existing same-day files. That mismatch can lead users or orchestrators to invoke the skill without realizing it performs persistent filesystem writes, increasing risk of unintended data loss or writing into sensitive locations.

Missing User Warnings

Medium

• Confidence: 93% confidence
• Finding: The document explicitly instructs users to schedule x-search runs and send results to external messaging platforms like Telegram, Slack, and Discord, but it does not warn that collected search content may be transmitted off-platform and into third-party services. In this skill context, that omission matters because scheduled automation increases the chance of repeated, unnoticed disclosure of monitored content, handles, topics, or derived summaries to destinations that may have different retention, access, and compliance properties.

Vague Triggers

Medium

• Confidence: 88% confidence
• Finding: The trigger language is broad enough to match ordinary conversation about X/Twitter, increasing the chance of unintended invocation. In an agent environment, accidental activation can cause unnecessary network calls, use of API credentials, and file output behavior without a clear user request to run the skill.

Self-Modification

High

• Category: Rogue Agent

Content

`--output` saves the result to a Markdown file (stdout is preserved):
- Pass a **directory** → auto-named file: `karpathy-2026-03-22.md`, `trends-AI-2026-03-22.md`, etc.
- Pass a **full path** → saved directly to that path
- Same-day runs overwrite the existing file (no duplicate accumulation)

`--progress-only` suppresses full Markdown on stdout — prints one summary line instead. Use with `--output` in automated pipelines (watchlist, cron) to avoid injecting large outputs into the agent context.

• Confidence: 94% confidence
• Finding: overwrite the existing file

Session Persistence

Medium

• Category: Rogue Agent

Content

`--output` saves the result to a Markdown file (stdout is preserved):
- Pass a **directory** → auto-named file: `karpathy-2026-03-22.md`, `trends-AI-2026-03-22.md`, etc.
- Pass a **full path** → saved directly to that path
- Same-day runs overwrite the existing file (no duplicate accumulation)

`--progress-only` suppresses full Markdown on stdout — prints one summary line instead. Use with `--output` in automated pipelines (watchlist, cron) to avoid injecting large outputs into the agent context.

• Confidence: 89% confidence
• Finding: write the existing file (no duplicate accumulation) `--progress-only` suppresses full Markdown on stdout — prints one summary line instead. Use with `--output` in automated pipelines (watchlist, cron

VirusTotal

64/64 vendors flagged this skill as clean.