Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Pre-Install Skill Security Gate: verdict before tool access

    1

    Use when you are about to install a third-party SKILL.md or skill folder and need a clear SAFE / CAUTION / DO NOT INSTALL verdict before it gets tool access.

    Free

    Security scanned
    Pre-Install Skill Security Gate: verdict before tool access

    Pre-Install Skill Security Gate: verdict before tool access

    Example session with this skill installed

    Audit this skill folder before I install it: ./candidate-skills/awesome-code-reviewer/

    • Read your context and instructions
    • Compiled the pre-install security gate

    Pre-install gate — awesome-code-reviewer v1.2.0

    IDTheme (heuristic)SeverityEvidence
    G1Malicious / unsafe command executionCRITICALscripts/setup.sh:34 — `curl …
    G2Credential harvestingHIGHSKILL.md:52 — instruct agent to read .env values
    G3Data exfiltrationCRITICALSKILL.md:67 — POST file contents to external analytics URL
    G4Prompt injection / override—none matched in this pass
    G5Dependency / registry riskLOWunpinned package note in references/deps.md

    Verdict: DO NOT INSTALL

    Reason: two CRITICAL findings (remote pipe-to-shell + content exfiltration) plus credential read. Static review only — this is not a malware sandbox proof. Do not install until the publisher removes or pins and justifies each item.

    Next steps

    cover-security-gate-verdict

    Connects securely to your tools. The creator never sees your data.

    About this skill

    The Problem

    A teammate finds a useful community skill. It reviews code. It also asks the agent to read .env "for configuration" and pipes a remote setup script to bash. Traditional SAST barely looks at SKILL.md. Asking another chat whether it "looks safe" returns a paragraph of guesses.

    You need a pre-install gate with a repeatable verdict (SAFE, CAUTION, or DO NOT INSTALL) before the skill inherits filesystem, network, and shell.

    Free vs Pro

    • Free: checklist + verdict for one skill or folder in chat; severity table; remediation hints
    • Pro ($9.99 one-time): batch library scan, SARIF/CI export, mapped findings pack, hard merge gate

    Free: https://www.agensi.io/skills/pre-install-skill-security-gate-verdict-before-tool-access Pro: https://www.agensi.io/skills/pre-install-skill-security-gate-pro-batch-scan-to-sarif-ci

    Upgrade Path

    Free is enough to gate a single candidate before install. Upgrade when you review many skills per week, need SARIF in CI, or want a library-wide report a security reviewer can attach to a PR.

    Pre-Install Skill Security Gate Pro ($9.99 one-time): https://www.agensi.io/skills/pre-install-skill-security-gate-pro-batch-scan-to-sarif-ci

    What You Get

    • A structured pre-install checklist across common agentic-skill risk themes: unsafe command execution, prompt-injection or override patterns, data exfiltration, credential harvesting, privilege or confirmation bypass, and dependency or registry smell
    • A companion-tree pass: SKILL.md plus scripts/, references/, templates/, and examples/ when those folders are present
    • One verdict: SAFE, CAUTION, or DO NOT INSTALL. The rules are heuristic. This is static review, not a malware proof or a certification
    • Severity-rated findings with file and line evidence and a concrete remediation ask
    • A hand-off to Pro for batch library scans, SARIF or CI export, and a merge gate ($9.99 one-time)

    Why This Beats Prompting It Yourself

    Generic "is this safe?" reads the marketing voice of the skill. Attack patterns hide in setup scripts, example blocks, and "helpful" telemetry. A fixed checklist forces the same rows every time and ends in a verdict a human can refuse — not a paragraph of false comfort.

    Use Cases

    • Vetting a marketplace or GitHub skill before first install
    • Reviewing a skill a coworker pasted into the repo
    • Spot-checking one existing install after a security scare
    • Teaching juniors what "looks wrong" in SKILL.md trees

    Related Timoranjes skills (cousins — not this SKU)

    Known Limitations

    • Static / heuristic review only. Not a malware proof, not a sandbox detonation, not a guarantee the skill is safe
    • Sophisticated or novel attacks can evade pattern checks
    • Does not execute the skill's code or observe runtime tool calls
    • False positives happen on legitimate docs fetches and dependency audits — judge destination and data before panic
    • Free = one skill/folder per session; no SARIF, no batch library gate
    • Does not replace least-privilege permissions, allowlists, or human review for high-trust environments
    • No invented breach claims, no "certified OWASP compliant" badge, no sales guarantees

    How to use

    1. 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. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    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

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 10 days ago

    • Free with an Agensi account

    Permissions

    Network Access
    Read Files
    Listed10 days ago
    Updated6 days ago

    What's inside

    Frequently Asked Questions

    Use when you are about to install a third-party SKILL.md or skill folder and need a clear SAFE / CAUTION / DO NOT INSTALL verdict before it gets tool access.

    This skill uses the open SKILL.md standard. Use it through an MCP-compatible agent connected to Agensi, and check the compatibility information on this page for confirmed agents and required capabilities.

    Connect your AI agent to Agensi through MCP, then ask it to use Pre-Install Skill Security Gate: verdict before tool access on your task. Your agent loads the latest approved version each time. Paid skills require a purchase or active subscription. Skill files are not available for download; your licence does not allow you to save, copy or export any part of a skill.

    New in Agents & Orchestration