Back to skill

Security audit

GitHub-README Hero.skill

Security checks for vulnerabilities and agentic risk

Overview

This skill only gives Markdown/README presentation guidance and does not request sensitive access, code execution, persistence, or hidden data handling.

This appears safe to install for README and Markdown presentation work. Be aware that implicit invocation is enabled, so review proposed edits before applying them if you need exact wording or structure preserved.

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
  • System Prompt LeakageDirect Leakage, Indirect Extraction, Tool-Based Exfiltration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (5)

Direct Prompt Extraction

High
Category
System Prompt Leakage
Confidence
85% confidence
Finding

Skill contains instructions that could directly expose system prompts, internal rules, or hidden instructions to users or external parties.

Content

Scanner excerpt · SKILL.md (reported line 163)May include surrounding context.

md
- Keep the first screen readable in raw Markdown and rendered view
- Make the page look designed, not overdesigned

## Output Rules

When you edit:
- keep the page functional as Markdown

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The default prompt uses a very generic invocation phrase that can match broad user requests about README editing without clear boundaries on when this skill should activate. In combination with implicit invocation, this increases the chance of unintended routing or over-application of the skill to content the user did not clearly intend to transform.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

Implicit invocation is enabled with no narrowing conditions, so the system may call this skill automatically based on loose semantic similarity rather than explicit user instruction. That can cause unexpected content rewriting, prompt-surface expansion, and accidental invocation in contexts where preserving exact text or structure matters.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
80% confidence
Finding

The badge text 'Chinese-First Friendly' expresses a language/locale orientation in the skill's natural-language description. Because the file does not clearly frame this as optional user preference or a region-specific constraint, it can be interpreted as a locale policy preference without explicit opt-in.

Content

No source excerpt is available for this finding.

Vague Triggers

Low
Category
Not specified by scanner
Confidence
85% confidence
Finding

The listed trigger phrases such as "Keep my intro, but make the top prettier" and "Why does their README look so much better than mine?" are relatively broad natural-language requests that could match general writing or design assistance. While the file includes some scope limits elsewhere, this trigger section does not provide negative examples or tighter constraints to reduce unintended invocation.

Content

No source excerpt is available for this finding.