CLI Help Writer

• Outcome: Pass
• Latest audit: Sep 16, 2026 · 4:36 AM UTC
• Version: 1.0.2

Overview

This is a straightforward CLI help-text writing skill with no executable code, persistence, network access, or hidden high-impact behavior.

Reasonable to install if you want help drafting CLI --help or man-page style text. Review generated output before copying it into a project, especially inferred defaults and redacted credential examples.

A.I.G

Vulnerability Patterns

• Skill Instruction Hijacking: Alters the agent's session goals or safety constraints when the skill loads
• Agent Memory Poisoning: Writes attacker-controlled rules into memory that affect later sessions
• Remote Payload Retrieval and Execution: Fetches external code whose behavior can change after review
• Embedded Malicious Code: Ships malicious scripts inside the skill and executes them locally
• Unauthorized Access and Privilege Escalation: Obtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns

• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Rogue Agent: Self-Modification, Session Persistence
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Behavioral AST: exec() Call, eval() Call, Dynamic Import
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration

Findings (7)

Instruction Override

High

• Category: Prompt Injection
• Confidence: 70% confidence

Finding

This pattern attempts to override system instructions or ignore safety constraints. Without LLM analysis, manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 89)May include surrounding context.

• Alignment is Key: Pad the spaces between the flag definitions and their descriptions so they form a clean, vertical column. Example:

  text

  -p, --port <number>      Port to listen on
  -d, --debug              Enable debug mode
  

• Infer Missing Info: If the user mentions "needs a port", invent a standard flag like -p, --port <number> with a reasonable default (e.g., 8080).
• Terminal Realism: Do not use bold (**) or italics (*) inside the text code block, as standard terminals do not render Markdown. Use uppercase letters for headers (e.g., OPTIONS:).

Session Persistence

Medium

• Category: Rogue Agent
• Confidence: 60% confidence

Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · SKILL.md (reported line 3)May include surrounding context.

md

---
name: cli-help-writer
description: "Generate standard, beautifully formatted `--help` documentation for Command Line Interface (CLI) tools based on raw arguments, flags, and descriptions. Triggers when the user asks to write help text, format CLI options, or generate a man page / terminal output text."
---

# CLI Help Writer Skill

Vague Triggers

Medium

• Category: Not specified by scanner
• Confidence: 92% confidence

Finding

The activation text says the skill triggers when the user asks to "write help text" or "format CLI options," which are broad natural-language phrases that could match many ordinary requests beyond an intended narrow invocation context. The description does not provide explicit constraints, exclusions, or negative examples to clarify when the skill should not activate.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

• Category: Not specified by scanner
• Confidence: 89% confidence

Finding

Natural-language policy checks apply to all file types. These evals require fixed-language outputs based on test definitions rather than an explicit user language preference mechanism or a documented, justified locale constraint, which can conflict with language-choice policies.

Content

No source excerpt is available for this finding.

Vague Triggers

Low

• Category: Not specified by scanner
• Confidence: 78% confidence

Finding

This is a manifest-style JSON file, so vague trigger analysis applies. The prompt text is a generic writing request and does not include any narrowing activation guidance, negative examples, or scope constraints, making invocation boundaries ambiguous if reused as skill-triggering language.

Content

No source excerpt is available for this finding.

Vague Triggers

Low

• Category: Not specified by scanner
• Confidence: 80% confidence

Finding

For manifest files, broad natural-language activation wording can cause unintended matching. This phrase does not define when the skill should activate versus when a general documentation or writing skill should respond.

Content

No source excerpt is available for this finding.

Vague Triggers

Low

• Category: Not specified by scanner
• Confidence: 76% confidence

Finding

In a JSON manifest/evals file, generic imperative phrases can be too vague if they reflect activation language. Here, the wording is broad and lacks any negative examples or explicit boundaries that distinguish this skill from general-purpose writing assistance.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.