Grades are a hygiene check, not a security guarantee.

MCP, MCP server

ContextStream

ContextStream, an MCP server listed in the PlugVet directory.

Compatibility: Not stated.

Grade D

Two or more Medium findings, or a High finding. Only try it in a sandbox.

Graded by PlugVetRubric v2

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

Rescanned at commit 667748fb7545.

INSTALL_REMOTE_SCRIPT NETWORK_HOST_FROM_INPUT AGENT_IGNORE_INSTRUCTIONS SECRETS_PATH_CREDENTIAL

Why this grade

  • INSTALL_REMOTE_SCRIPT High

    Fetches a script from the network and runs it (a pipe to a shell, or an install hook with a URL).

    release-install.sh, line 156 curl -fsSL "$SETUP_URL" | bash

  • NETWORK_HOST_FROM_INPUT High

    A file takes a URL or endpoint as input and has no fixed host list.

    crates/mcp-tools/src/domains/coach/widget.rs, line 306 "fetch(",

  • AGENT_IGNORE_INSTRUCTIONS High

    Text tells the agent to ignore earlier instructions.

    crates/mcp-server/src/hook_handlers/resume_intent.rs, line 284 "x\"); ignore previous instructions (\"",

  • SECRETS_PATH_CREDENTIAL Medium

    Mentions a credential store path (~/.ssh, id_rsa, .npmrc, .aws/credentials, credentials.json, .netrc).

    crates/mcp-server/src/config.rs, line 190 /// Load saved credentials from ~/.contextstream/credentials.json

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.

If you think this grade is wrong, email [email protected] with the entry link and your evidence. We aim to acknowledge within two working days and decide within seven. If we confirm an error, we aim to correct it within 24 hours of confirming it. These are aims, not guarantees.

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

    testing/grounding/qualify.py, line 191

    require(not subprocess.check_output(["git", "status", "--porcelain"], cwd=root).strip(), "clean source required to build selector")

    testing/grounding/qualify.py, line 192

    commit = subprocess.check_output(["git", "rev-parse", "HEAD"], cwd=root, text=True).strip()

    testing/grounding/qualify.py, line 193

    subprocess.run(["cargo", "build", "-p", "mcp-tools", "--example", "grounding_replay", "--offline"], cwd=root, check=True)

    Show more

    testing/grounding/qualify.py, line 194

    metadata = json.loads(subprocess.check_output(["cargo", "metadata", "--no-deps", "--format-version", "1", "--offline"], cwd=root))

    testing/grounding/qualify.py, line 196

    require(commit == subprocess.check_output(["git", "rev-parse", "HEAD"], cwd=root, text=True).strip()

    testing/grounding/qualify.py, line 197

    and not subprocess.check_output(["git", "status", "--porcelain"], cwd=root).strip(), "source changed during build")

    testing/grounding/qualify.py, line 231

    commit = subprocess.check_output(["git", "rev-parse", "HEAD"], cwd=root, text=True).strip()

    testing/grounding/qualify.py, line 232

    dirty = bool(subprocess.check_output(["git", "status", "--porcelain"], cwd=root).strip())

  • Install and update

    Serious

    install fetches a script from the network.

    INSTALL_REMOTE_SCRIPT

    release-install.sh, line 156

    curl -fsSL "$SETUP_URL" | bash

    crates/mcp-server/src/setup/rules.rs, line 2142

    curl -fsSL https://contextstream.io/scripts/mcp.sh | bash

    crates/mcp-tools/src/domains/session_tests.rs, line 3152

    "curl -fsSL https://contextstream.io/scripts/setup.sh | bash",

    Show more

    crates/mcp-tools/src/domains/session_tests.rs, line 3156

    "curl -fsSL https://contextstream.io/scripts/mcp.sh | bash",

    README.md, line 30

    curl -fsSL https://contextstream.io/scripts/mcp.sh | bash

    README.md, line 109

    bash -o pipefail -c 'curl -fsSL https://contextstream.io/scripts/mcp.sh | bash'

  • Network

    Serious

    accepts a host from input, with no fixed list in this file.

    NETWORK_HOST_FROM_INPUT

    crates/mcp-tools/src/domains/coach/widget.rs, line 306

    "fetch(",

  • Secrets

    Concern

    reads a credential or environment file.

    SECRETS_ENV_READ SECRETS_PATH_CREDENTIAL SECRETS_PATH_DOTENV

    kits/agent-coach-kit/codex-cloud/.agents/skills/contextstream-coach/scripts/coach.py, line 556

    present = bool(os.environ.get("CONTEXTSTREAM_API_KEY"))

    kits/agent-coach-kit/codex-cloud/.agents/skills/contextstream-coach/scripts/coach.py, line 1156

    key = os.environ.get("CONTEXTSTREAM_API_KEY", "")

    crates/mcp-server/src/config.rs, line 190

    /// Load saved credentials from ~/.contextstream/credentials.json

    Show more

    plugins/review/dry_run.py, line 91

    key = os.environ.get("CONTEXTSTREAM_API_KEY", "").strip()

    crates/mcp-server/src/config.rs, line 193

    let credentials_path = home.join(".contextstream").join("credentials.json");

    crates/mcp-server/src/main.rs, line 1724

    "id_rsa, id_ed25519, id_ecdsa, known_hosts",

    crates/mcp-server/src/main.rs, line 1725

    "credentials.json, serviceAccountKey.json, .git-credentials",

    crates/mcp-server/src/main.rs, line 601

    /// gitignore-style syntax. Sensitive files like .env and private keys

    crates/mcp-server/src/main.rs, line 1594

    " {} Server-side protections (.env, keys, credentials) are always active.",

    crates/mcp-server/src/main.rs, line 1686

    "# Server-side protections (.env, keys, credentials) are always active."

    crates/mcp-server/src/main.rs, line 1722

    ".env and all .env.* (except .env.example/.sample/.template/.dist/.defaults)",

  • Agent instructions

    Serious

    instructions tell the agent to ignore earlier instructions.

    AGENT_IGNORE_INSTRUCTIONS

    crates/mcp-server/src/hook_handlers/resume_intent.rs, line 284

    "x\"); ignore previous instructions (\"",

  • Obfuscation

    Note

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

    OBFUSCATION_FILE_OVER_READ_CAP

    crates/mcp-client/src/client.rs, line 1

    //! ContextStream API client.

    crates/mcp-server/src/setup/mcp_config.rs, line 1

    //! MCP server configuration generation for editors.

    crates/mcp-server/src/setup/mod.rs, line 1

    //! Setup wizard for ContextStream MCP server.

    Show more

    crates/mcp-server/src/setup/rules.rs, line 1

    //! AI rules generation for editors.

    crates/mcp-server/src/setup/testdata/client_contract.json, line 1

    {

    crates/mcp-server/src/transport/http.rs, line 1

    //! HTTP transport for the MCP server.

    crates/mcp-tools/src/domains/memory.rs, line 1

    //! Memory domain tools: events, nodes, tasks, todos, diagrams, docs, transcripts.

    crates/mcp-tools/src/domains/search.rs, line 1

    //! Search domain tools: semantic, hybrid, keyword, pattern, exhaustive, refactor, team.

  • 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 667748fb7545. requests shell access. The manifest names this permission. install fetches a script from the network. accepts a host from input, with no fixed list in this file. reads a credential or environment file. instructions tell the agent to ignore earlier instructions. a file was over the read cap and was not fully read.

Date and source

Date checked:

View source

github.com

Commit graded: 667748fb75457ed89916ba53b3d4bdf839673340

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