More screenshots
Works with the AI Agents you already use
Skill Safety
Run the safety pre-check a marketplace reviewer would run on a skill: 8 checks and a 0-100 score, then explain and fix each finding.
$6
About this skill
The problem. A skill is rejected for safety and the reason is unclear, or you ship a script and are not sure how it looks to a reviewer.
What it does. A read-only script lays out a marketplace-style pre-check of a skill folder: 8 checks and a 0-100 score. 1 file structure, 2 file types (no binaries, archives or symlinks), 3 dangerous commands (curl | sh, sudo, rm -rf /, chmod 777), 4 secrets (keys and tokens), 5 environment harvesting (reading the whole environment or credential-like variables; combined with network use it fails), 6 network access (allowed when SKILL.md says so and names the host), 7 obfuscation and hidden characters (base64 that decodes to text, decode-and-run, zero-width and text-direction characters), 8 prompt injection (phrases that tell the agent to ignore its rules or hide something from the user).
You get: a verdict (BLOCKED, FIX FIRST or READY), the score, each finding with file:line, and how to fix it: remove the cause, declare the network use and host in SKILL.md, rotate an exposed key, and read docs that only mention a risky phrase to decide whether it is a warning or an order. It also lists likely false positives, such as a security skill that quotes attack examples.
Rules built in: never rename or reshape a risky command to slip past the check; the score is our estimate, not the marketplace's; a READY result does not guarantee approval. If the script is not available, the skill does the 8 checks by reading and marks them "done by reading".
Part of the Skill Publisher QA kit. Works alone.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Security.
The score is an estimate; it cannot know the marketplace's exact rules or threshold; pattern checks can flag harmless text such as security examples in docs.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed today
- Software
Python 3
Permissions
File Scopes
Runs agensi_precheck.py with python3 -I. It reads files only: no network, no environment variables, nothing from the skill is run.