Grades are a hygiene check, not a security guarantee.

MCP, MCP server

Cost Guard Mcp

Cost Guard Mcp, an MCP server listed in the PlugVet directory.

Compatibility: Not stated.

Grade B

Minor issues only, no Medium or High. Worth a quick look at the findings below before installing.

Graded by PlugVetRubric v2

Not a guarantee. A hygiene check of the source at the commit shown, not a security audit.

NETWORK_NO_FIXED_HOST SECRETS_PATH_DOTENV PERMISSION_SHELL_DECLARED SECRETS_ENV_READ OBFUSCATION_FILE_OVER_READ_CAP

Why this grade

  • NETWORK_NO_FIXED_HOST Low

    Makes a network call whose host is not written as a fixed address.

    DECISIONS.md, line 1120 fetch (the `max_rows + 1` sentinel -> `row_cap_exceeded`, `executed: true`) and the 8 MiB cap

  • SECRETS_PATH_DOTENV Low

    Mentions a .env file.

    .claude/skills/verify/SKILL.md, line 15 Hard rules: never read/source/copy `.env` or a real credential (dummy values only, generated into a

  • PERMISSION_SHELL_DECLARED Note

    Runs shell commands, and the manifest or description says so.

    .claude/skills/verify/scripts/stdio_rpc.py, line 64 self.proc = subprocess.Popen( # noqa: S603 - the command IS the thing under test

  • SECRETS_ENV_READ Note

    Reads an environment variable whose name contains KEY, SECRET, TOKEN, PASSWORD or CREDENTIAL.

    src/cost_guard_mcp/config.py, line 46 creds = os.environ.get("GOOGLE_APPLICATION_CREDENTIALS")

  • OBFUSCATION_FILE_OVER_READ_CAP Note

    A file was over the read cap and was not fully read.

    uv.lock, line 1 version = 1

Any High, or two or more Medium findings, give D. One Medium gives C. Low findings or a security Note give B. Each code names a finding behind the letter.

Grades are a hygiene check of the source at the commit shown, on the date checked, not a security audit or a guarantee. The free directory grades are automated. Before release, a PlugVet engineer reviews every paid report and any grade we hold back, including all F grades, against the source code and the evidence we quote.

What was checked

Plain notes from the hygiene checks.

  • Permissions

    Note

    requests shell access. The manifest names this permission.

    PERMISSION_SHELL_DECLARED

    .claude/skills/verify/scripts/stdio_rpc.py, line 64

    self.proc = subprocess.Popen( # noqa: S603 - the command IS the thing under test

  • Install and update

    Pass

    No install-time script found, and dependencies are pinned or absent.

  • Network

    Concern

    can reach a host that is not written as a fixed address.

    NETWORK_NO_FIXED_HOST

    DECISIONS.md, line 1120

    fetch (the `max_rows + 1` sentinel -> `row_cap_exceeded`, `executed: true`) and the 8 MiB cap

    changelog/v0.6.0.md, line 119

    `max_results`/`page_size`; the Snowflake post-poll fetch (and Databricks' post-join fetch) now

    src/cost_guard_mcp/engines/databricks.py, line 52

    # request is a requests.post without a timeout. Wrapping _connect in a thread+join watchdog

  • Secrets

    Concern

    reads a credential or environment file.

    SECRETS_ENV_READ SECRETS_PATH_DOTENV

    src/cost_guard_mcp/config.py, line 46

    creds = os.environ.get("GOOGLE_APPLICATION_CREDENTIALS")

    src/cost_guard_mcp/config.py, line 142

    private_key_path = os.environ.get("SNOWFLAKE_PRIVATE_KEY_PATH")

    .claude/skills/verify/SKILL.md, line 15

    Hard rules: never read/source/copy `.env` or a real credential (dummy values only, generated into a

    Show more

    src/cost_guard_mcp/config.py, line 143

    password = os.environ.get("SNOWFLAKE_PASSWORD")

    src/cost_guard_mcp/config.py, line 162

    private_key_passphrase=os.environ.get("SNOWFLAKE_PRIVATE_KEY_PASSPHRASE") or None,

    src/cost_guard_mcp/config.py, line 187

    access_token = os.environ.get("DATABRICKS_TOKEN")

    src/cost_guard_mcp/config.py, line 189

    client_secret = os.environ.get("DATABRICKS_CLIENT_SECRET")

    .claude/skills/verify/SKILL.md, line 131

    git -C $W check-ignore -v .env .mcp.json service-account.json .cursor/mcp.json # read-only on the worktree

    .gitignore, line 3

    .venv/

    .gitignore, line 25

    # .env is ignored above; its per-environment variants must be too.

    .gitignore, line 27

    # ...except a committed example (placeholder values only), which .env.* would silently hide.

    .gitignore, line 28

    !.env.example

    DECISIONS.md, line 162

    a genuinely different risk profile than a local `.env` file.

    README.md, line 109

    Whatever MCP client/host you use (Claude Desktop, etc.) spawns this server as its own subprocess - it does **not** automatically inherit your shell's environment variables, even if

  • Agent instructions

    Pass

    No instruction to ignore earlier instructions or to hide a step.

  • Obfuscation

    Note

    a file was over the read cap and was not fully read.

    OBFUSCATION_FILE_OVER_READ_CAP

    uv.lock, line 1

    version = 1

  • Licence

    Pass

    A licence file is present (LICENSE).

  • Maintainer activity

    Pass

    The latest commit is within the last 365 days.

Get the full report on this tool, £49

Note

Hygiene check of commit c59fec58ada2. requests shell access. The manifest names this permission. can reach a host that is not written as a fixed address. reads a credential or environment file. a file was over the read cap and was not fully read.

Date and source

Date checked:

View source

github.com

Commit graded: c59fec58ada275b54f82198e0b765fe30c7c560b

This link goes to the original host. PlugVet does not host or check what you download from there, and installing it is your own decision and risk.

Read how this letter is chosen