Agent Skill Adoption Gatekeeper

    1

    Review a supplied Agent Skill before installation and produce a documented adopt, sandbox, revise, or decline decision with provenance, permissions, update, and approval controls.

    Secure checkout via Stripe

    0 installsSecurity scanned

    Works with the AI tools you already use

    CClaude CodeCCursorCCodex CLIGGitHub CopilotGGemini CLI+17 more

    See it in action

    You say

    Review this SKILL.md for a new repository-summarizer tool. It requests read access to all files and network egress to an analytics endpoint. Version is unpinned on GitHub.

    Your agent does

    ### Review Scope Artifact: repository-summarizer (unpinned GitHub reference). ### Capability Map - Read Files: Workspace scope (Affected: project source). Posture: Bounded. - Network Egress: Analytics endpoint. Posture: Untrusted. ### Disposition **Sandbox and verify.** Risk: Unpinned version allows stealth updates. Deny network egress by default.

    What you get

    Audit third-party MCP servers before local installation.Map agent permission requests to actual system risk levels.Identify unpinned dependencies and mutable instruction sources.Generate documented adoption records for engineering teams.

    About this skill

    What it does

    Turn a supplied SKILL.md, manifest, dependency list, file tree, publisher record, or release reference into a pre-installation decision packet. Create a provenance register, least-privilege capability map, external-instruction and dependency boundary review, recommended controls, and a clear disposition.

    Best for

    AI platform engineers, technical founders, AppSec reviewers, and teams adopting third-party Agent Skills across Claude Code, Codex, Cursor, OpenClaw, Gemini CLI, or other compatible runtimes.

    What makes it different

    This is not a generic security reviewer or malware scanner. It structures the decision a team must make before installation: which evidence is present, what is unknown, what permissions are acceptable, which controls are needed, and whether the package should be adopted, sandboxed, revised, or declined.

    Known limitations

    It performs a static review of material the buyer supplies. It does not execute code, inspect a live environment, access accounts, verify publisher identity, scan for malware, sign packages, or certify security, compliance, or safety.

    How to install

    Drop the file into your AI Agent. Works with Claude, Cursor, ChatGPT, and 20+ more.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill passes our trust checks below.

    Security scanned

    Passed our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    30-day refund

    Not a fit? Get your money back

    Trust & safety

    Security scanned

    Verified clean today

    Listedtoday

    Frequently Asked Questions