@vibo-dev/skill-injection-scanner 1.1.2 β†’ 1.1.4

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/README.md CHANGED
@@ -1,5 +1,11 @@
1
1
  # πŸ” Skill Injection Scanner
2
2
 
3
+ [![SkillQA CI](https://github.com/vnbochkarev-netizen/skill-injection-scanner/actions/workflows/skillqa-ci.yml/badge.svg)](https://github.com/vnbochkarev-netizen/skill-injection-scanner/actions/workflows/skillqa-ci.yml)
4
+ [![License: MIT](https://img.shields.io/badge/license-MIT-blue.svg)](LICENSE)
5
+ [![npm](https://img.shields.io/npm/v/@vibo-dev/skill-injection-scanner)](https://www.npmjs.com/package/@vibo-dev/skill-injection-scanner)
6
+
7
+ <p align="center"><img src="assets/banner.svg" alt="Skill Injection Scanner" width="100%"></p>
8
+
3
9
  **Find hidden instructions and prompt-injection patterns inside your agent's skill files β€” before they find you.**
4
10
 
5
11
  Skill marketplaces are booming (ClawHub, n8n, OpenClaw…). So is the dark side:
@@ -19,8 +25,11 @@ obfuscation, remote-instruction fetches, and manipulation tricks β€” in **Englis
19
25
  ## Install
20
26
 
21
27
  ```bash
22
- # npm (any OS with Python 3.8+)
23
- npx -y @vibo-dev/skill-injection-scanner --skills ~/.openclaw/skills
28
+ # direct (recommended β€” runs the reviewed source)
29
+ python3 scanner.py --skills ~/.openclaw/skills
30
+
31
+ # npm β€” pin the reviewed release (any OS with Python 3.8+)
32
+ npx -y @vibo-dev/skill-injection-scanner@1.1.4 --skills ~/.openclaw/skills
24
33
 
25
34
  # ClawHub / OpenClaw registry: install "skill-injection-scanner"
26
35
  # GitHub: clone this repo
package/SKILL.md CHANGED
@@ -1,7 +1,7 @@
1
1
  ---
2
2
  name: skill-injection-scanner
3
3
  description: "Scan agent skill files for hidden instructions and prompt-injection patterns (EN/RU) before a poisoned skill rewrites your agent. 19 rules, zero dependencies. Use ONLY with the user's explicit consent: tell the user which skills folder will be scanned β€” findings are printed to stdout locally."
4
- version: 1.1.2
4
+ version: 1.1.4
5
5
  tools: [python]
6
6
  license: MIT
7
7
  ---
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@vibo-dev/skill-injection-scanner",
3
- "version": "1.1.2",
3
+ "version": "1.1.4",
4
4
  "description": "Find hidden instructions and prompt-injection patterns inside agent skill files (EN/RU, 19 rules). Local-first, zero dependencies, Python 3.8+.",
5
5
  "license": "MIT",
6
6
  "author": "Viacheslav Bochkarev <vnbochkarev@gmail.com>",
package/scanner.py CHANGED
@@ -33,7 +33,7 @@ import os
33
33
  import re
34
34
  import sys
35
35
 
36
- __version__ = "1.1.2"
36
+ __version__ = "1.1.3"
37
37
 
38
38
  RULES = [
39
39
  # (name, regex, severity, note)