Grades are a hygiene check, not a security guarantee.

MCP, MCP server

atlas

Open-source local-first cognitive memory — alpha. Implements AGM-compliant belief revision on a property graph.

Compatibility: Not stated.

Grade D

Two or more Medium findings, or a High. 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.

INSTALL_REMOTE_SCRIPT

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).

    laptop-setup/install.sh, line 351 curl -fsSL https://raw.githubusercontent.com/mediar-ai/screenpipe/main/install.sh | bash 2>&1 | tee -a "$LOG_FILE" || true

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

    Concern

    can delete files.

    PERMISSION_DELETE_FILES

    atlas_core/cli.py, line 212

    completed = subprocess.run([str(script), *args.demo_args]) # noqa: S603

    atlas_core/api/auth.py, line 19

    subprocess.run(

    benchmarks/business_mem_bench/adapters/atlas_system.py, line 70

    shutil.rmtree(self._data_dir, ignore_errors=True)

    Show more

    benchmarks/business_mem_bench/adapters/atlas_system.py, line 533

    shutil.rmtree(self._data_dir, ignore_errors=True)

    integrations/hermes-atlas/atlas/cognitive_client.py, line 55

    self._process: subprocess.Popen[Any] | None = None

    integrations/hermes-atlas/atlas/cognitive_client.py, line 245

    self._process = subprocess.Popen(command, **options)

    scripts/doctor.py, line 80

    out = subprocess.run(

    scripts/doctor.py, line 98

    out = subprocess.run(

    scripts/doctor.py, line 214

    out = subprocess.run(

    scripts/qa_site.py, line 91

    out = subprocess.run(

    scripts/run_bmb.py, line 58

    shutil.rmtree(args.corpus)

  • Install and update

    Serious

    install fetches a script from the network.

    INSTALL_REMOTE_SCRIPT

    laptop-setup/install.sh, line 351

    curl -fsSL https://raw.githubusercontent.com/mediar-ai/screenpipe/main/install.sh | bash 2>&1 | tee -a "$LOG_FILE" || true

  • Network

    Concern

    can reach a host that is not written as a fixed address. can reach raw.githubusercontent.com. The description does not name that host.

    NETWORK_NO_FIXED_HOST NETWORK_UNNAMED_HOST

    demo.sh, line 79

    if curl -sf http://localhost:7474 >/dev/null 2>&1; then

    demo.sh, line 85

    curl -sf http://localhost:7474 >/dev/null 2>&1 && break

    demo.sh, line 88

    if curl -sf http://localhost:7474 >/dev/null 2>&1; then

    Show more

    atlas_core/cli.py, line 5

    Cypher query, or a curl call, even though the underlying functions were all

    atlas_core/api/http_server.py, line 4

    web dashboard, programmatic curl access, integration tests).

    atlas_core/retrieval/vault_search.py, line 76

    response = httpx.post(

    atlas_core/retrieval/vault_search.py, line 113

    response = httpx.get(

    integrations/cognitive-service/clients/node-client.mjs, line 42

    const response = await fetch(`${baseUrl.replace(/\/$/, "")}${spec.path}`, {

    integrations/cognitive-service/clients/python_client.py, line 65

    with urllib.request.urlopen(request, timeout=step.get("timeout_seconds", 10)) as response:

    integrations/hermes-atlas/atlas/cognitive_client.py, line 327

    with urllib.request.urlopen(request, timeout=self.timeout_seconds) as response:

    integrations/openclaw-atlas/src/cognitive-client.ts, line 170

    response = await fetch(`${this.baseUrl}${path}`, {

    laptop-setup/install.sh, line 156

    if curl -sf http://localhost:7474 >/dev/null 2>&1; then

    laptop-setup/install.sh, line 321

    if curl -sf http://127.0.0.1:8384 >/dev/null 2>&1; then

    laptop-setup/install.sh, line 458

    if curl -sf http://127.0.0.1:8384 >/dev/null 2>&1; then

    scripts/demo_loop.py, line 98

    print(f" {DIM}curl http://localhost:7474 # should return 200{RESET}")

  • Secrets

    Concern

    reads a credential or environment file.

    SECRETS_ENV_READ SECRETS_PATH_DOTENV

    atlas_core/graphiti.py, line 31

    api_key = os.environ.get("ANTHROPIC_API_KEY")

    atlas_core/adapters/claude_code.py, line 67

    password = os.environ.get("ATLAS_NEO4J_PASSWORD", "atlasdev")

    atlas_core/api/api_app.py, line 43

    os.environ.get("NEO4J_PASSWORD", "atlasdev"),

    Show more

    atlas_core/api/auth.py, line 55

    configured = os.environ.get("ATLAS_HTTP_TOKEN")

    atlas_core/ingestion/extractors/llm_base.py, line 94

    # .env file work for both CLI entrypoints and direct Python API use.

    atlas_core/ingestion/extractors/llm_base.py, line 98

    api_key = os.environ.get("ANTHROPIC_API_KEY")

    atlas_core/lineage/extractor.py, line 87

    api_key = os.environ.get("ANTHROPIC_API_KEY")

    atlas_core/resolution/llm_fallback.py, line 135

    api_key = os.environ.get("ANTHROPIC_API_KEY")

    atlas_core/working/auto_summarizer.py, line 69

    api_key = os.environ.get("ANTHROPIC_API_KEY")

    scripts/doctor.py, line 187

    env_path = _REPO_ROOT / ".env"

    scripts/doctor.py, line 192

    ".env not present (only needed for live ingestion with API keys)",

    scripts/doctor.py, line 247

    (".env file (optional)", check_env_file),

    .env.example, line 2

    # Copy to .env and fill in the secrets that aren't already in 1Password.

    .gitignore, line 29

    .atlas-laptop-setup.state

    .gitignore, line 30

    .env

    docs/LAUNCH_BACKLOG.md, line 47

    Bolt port, APOC version, `~/.atlas` writability, `.env`

  • Agent instructions

    Pass

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

  • Obfuscation

    Pass

    No encoded or minified payload found.

  • Licence

    Pass

    A licence file is present (LICENSE).

  • Maintainer activity

    Pass

    The latest commit is within the last 365 days.

Note

Hygiene check of commit 229df544944f. can delete files. install fetches a script from the network. can reach a host that is not written as a fixed address. can reach raw.githubusercontent.com. The description does not name that host. reads a credential or environment file.

Date and source

Date checked:

View source

github.com

Commit graded: 229df544944ffb1dadf4d155bcb41334e850da0c

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