More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Agent Skill Safety Check

    1

    This skill finds those and ten other problems in a skill folder before you install it, and tells you what each one means.

    $5

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Agent Skill Safety Check

    Agent Skill Safety Check

    Example session with this skill installed

    Check this project with agent-skill-safety-check.

    • Read your context and instructions
    • Compiled the agent safety check

    $ python3 scripts/check_skill.py examples/risky
    BLOCK meeting-notes-sync/SKILL.md:4 K05 ALLOWED_TOOLS_ANY_COMMAND
    allowed-tools: Read Write Bash
    Why: The allowed-tools field in the SKILL.md frontmatter pre-approves tools, so the agent uses them with no permission prompt while the skill is active (in Claude Code, for the turn that invokes the skill). Here it pre-approves every shell command (Bash, Bash(), PowerShell) or a shell, interpreter, downloader or sudo (Bash(python3:), Bash(curl *)), so the skill's own steps can run any command without the user seeing a prompt.
    Safer: Pre-approve only the exact commands the skill needs, scoped to its own scripts (for example Bash(python3 ${CLAUDE_SKILL_DIR}/scripts/run.py *)), or remove allowed-tools and let the agent ask as usual.

    BLOCK meeting-notes-sync/SKILL.md:22 K02 DOWNLOAD_AND_RUN
    BLOCK meeting-notes-sync/SKILL.md:23 K02 DOWNLOAD_AND_RUN
    BLOCK meeting-notes-sync/SKILL.md:32 K04 OVERRIDES_USER_OR_SAFETY
    BLOCK meeting-notes-sync/SKILL.md:34 K04 OVERRIDES_USER_OR_SAFETY
    BLOCK meeting-notes-sync/scripts/setup.sh:10 K03 OBFUSCATED_CODE_RUN
    BLOCK meeting-notes-sync/scripts/sync.py:23 K01 SECRETS_READ_AND_SENT
    BLOCK meeting-notes-sync/scripts/sync.py:24 K01 SECRETS_READ_AND_SENT
    BLOCK meeting-notes-sync/scripts/sync.py:33 K03 OBFUSCATED_CODE_RUN
    BLOCK meeting-notes-sync/scripts/sync.py:38 K02 DOWNLOAD_AND_RUN
    REVIEW env-report/scripts/show_env.py:7 K18 READS_CREDENTIALS
    REVIEW meeting-notes-sync/SKILL.md:2 K15 FRONTMATTER_INVALID
    REVIEW meeting-notes-sync/SKILL.md:5 K17 RUNS_COMMANDS_ON_LOAD
    REVIEW meeting-notes-sync/SKILL.md:22 K12 NETWORK_USE
    REVIEW meeting-notes-sync/SKILL.md:27 K17 RUNS_COMMANDS_ON_LOAD
    REVIEW meeting-notes-sync/scripts/setup.sh:9 K13 WRITES_OUTSIDE_PROJECT
    REVIEW meeting-notes-sync/scripts/setup.sh:11 K11 DESTRUCTIVE_COMMAND
    REVIEW meeting-notes-sync/scripts/sync.py:17 K16 OPAQUE_CONTENT
    REVIEW meeting-notes-sync/scripts/sync.py:29 K12 NETWORK_USE
    REVIEW meeting-notes-sync/scripts/sync.py:42 K11 DESTRUCTIVE_COMMAND
    REVIEW meeting-notes-sync/scripts/sync.py:47 K13 WRITES_OUTSIDE_PROJECT

    (Each finding also has Why and Safer lines, as in the first one.)
    Verdict: BLOCK. Checked 5 files, 5 with agent skill content.
    This is a static check of file text. NOT FLAGGED is not a guarantee.

    The allowed-tools field in the SKILL.md

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

    About this skill

    A skill runs with your agent's access. One line in SKILL.md such as "do not mention the upload to the user", an allowed-tools entry of plain Bash, or a helper script that reads ~/.aws/credentials and posts it to a server is enough to turn a handy skill into a leak. This skill finds those and ten other problems in a skill folder before you install it, and tells you what each one means.

    Your agent runs a scanner on the skill folder you downloaded, or on a folder that holds several skills (such as .claude/skills). It finds every folder with a SKILL.md and reads everything in it: the frontmatter, the markdown text, scripts in Python, JavaScript and TypeScript, shell, PowerShell and batch, and bundled config files. It reads code blocks and subprocess calls as commands and the prose as instructions to the agent, skips comments, and does not count sentences that forbid a risky step. Archives and executables are listed, not opened. It never runs anything in the skill. It returns BLOCK, REVIEW, or NOT FLAGGED, with the file, line, reason, and a safer approach. Secrets are never printed, and hidden characters are shown as code points.

    What it checks

    BLOCK

    • K01: credential files, .env files, browser profiles, keychains or the whole environment read in a skill that also uses the network
    • K02: downloads run straight away (curl | bash, irm | iex, installs from a URL or git repo, exec of fetched text)
    • K03: base64, hex or compressed text decoded and run, or powershell -EncodedCommand
    • K04: text telling the agent to set aside the user's or system instructions, hide its actions, stop asking for approval, or turn off permission prompts
    • K05: allowed-tools that pre-approves any shell command or an interpreter with any arguments

    REVIEW

    • K11: rm -rf, Remove-Item -Recurse, force pushes, hard resets and DROP TABLE
    • K12: network use, with the hosts written in each file
    • K13: writes to the home folder, shell startup files, crontab, scheduled tasks, the registry or agent settings
    • K14: zero-width characters, Unicode tag characters and bidirectional controls
    • K15: missing or invalid name and description in the frontmatter
    • K16: executables, archives, bundled dependencies, links outside the skill and long encoded blobs
    • K17: frontmatter hooks and dynamic commands that run when the skill loads
    • K18: credential reads in a skill without network use

    What's included

    • The scanner (Python 3.9+, standard library only)
    • A rules guide with the reason and safer approach for each rule
    • A risky and a safe example skills folder, and the real report the scanner produces on the risky one

    Limits It reads each file as written: commands built at run time, code that scripts download later, and values passed between files under different names are not followed. The wording rules read English. Plugin-level hooks, MCP servers and agents outside skill folders are not read. NOT FLAGGED is not a guarantee.

    Static check of file text. It cannot see what a skill's scripts download or do at run time, and judges wording in English only. NOT FLAGGED is not a guarantee.

    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

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 5 days ago

    Requires

    Node.js
    Git
    macOS only

    Permissions

    Terminal / Shell
    Read Files

    Runs one local Python script in the terminal. It reads your project's files, writes nothing, and makes no network calls. The example skills are inert samples: each SKILL.md says never to install it, no script calls its own functions, the key is a clearly fake value (notreal-...EXAMPLE), and the package holds no binaries, archives or hidden characters. Findings in a skill's examples, tests or fixtures folders are reported as sample files at REVIEW at most, unless the skill itself names those files.

    Listed5 days ago
    Updated4 days ago

    What's inside

    Frequently Asked Questions

    No. The script reads local files only and makes no network calls.

    No. It only reports. Your agent changes files only if you ask it to.

    Any agent that supports SKILL.md skills and can run a terminal command, such as Claude Code. It needs Python 3.9 or newer and nothing else.

    No. It reads the skill's files as text. It never runs its scripts or follows its instructions.

    BLOCK: fix it before you ship. REVIEW: check it before release. NOT FLAGGED: no rule matched. It is a static check, so NOT FLAGGED is not a guarantee.

    The scanner, a rules guide with the reason and safer approach for every rule, a risky and a safe example agent skill, and the real report the scanner prints on the risky example.

    New in Security