io.github.lonniev/personalbrain-mcp

What we checked

Six static checks, weighted by risk. Every result reflects only what could be observed in the published package and repository — never intent.

• Injection surface

  warn25% of grade

  Tool descriptions/manifest scanned for instruction-injection patterns (imperatives at the model, hidden text, 'ignore previous', data-exfil URLs).

  tool "brain_create_or_update_note" description: off-vendor URL in a description (…); capability text should not name where data goes. tool "brain_append_to_note" description: off-vendor URL in a description (…); capability text should not name where data goes.

• Supply chain

  fail25% of grade

  Package provenance: namespace verification, repo linkage, maintainer count, account age, postinstall scripts, typosquat distance.

  Python build script executes at build time: network call in build script

• Credential hygiene

  fail15% of grade

  How the server takes secrets (env vs plaintext config vs hardcoded); secrets appearing in tool schemas.

  tool "brain_check_balance" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_purchase_credits" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_check_payment" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_restore_credits" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_account_statement" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_account_statement_infographic" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_report_issue" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_get_patron_onboarding_status" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_receive_credentials" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_receive_credentials" has a secret-shaped input parameter "credential_card" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_forget_credentials" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_receive_patron_credentials" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_receive_patron_credentials" has a secret-shaped input parameter "credential_card" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_update_patron_credential" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_update_operator_credential" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_delete_patron_credential" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_delete_operator_credential" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_get_patron_credential_fields" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_receive_npub_proof" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_check_proof_status" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_set_pricing_model" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_reset_pricing_model" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_request_adoption" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_adoption_status" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_restore_neon_schema" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_check_price" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_mint_coupon" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_list_coupons" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_update_coupon" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_delete_coupon" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_redeem_coupon" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_list_my_coupons" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_forget_coupon" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_list_brains" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_get_brain" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_set_active_brain" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_get_brain_stats" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_create_thought" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_get_thought" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_get_thought_by_name" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_update_thought" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_delete_thought" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_search_thoughts" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_get_thought_graph" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_get_thought_graph_paginated" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_get_types" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_get_tags" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_create_link" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_update_link" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_get_link" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_delete_link" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_add_file_attachment" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_add_url_attachment" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_get_attachment" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_get_attachment_content" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_delete_attachment" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_list_attachments" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_get_note" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_create_or_update_note" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_append_to_note" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_get_modifications" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_brain_query" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_morph_thought" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_scan_orphans" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential). tool "brain_event_for_person" has a secret-shaped input parameter "dpop_token" — secrets must never be tool parameters (the model would supply/handle the credential).

• Permission scope

  pass15% of grade

  Declared tools vs. breadth (filesystem, network, exec); flags shell-exec and unbounded filesystem access.

  Network egress present (pyproject declares an HTTP client dependency) — expected for most servers and not a finding on its own. No shell-exec, unbounded/sensitive filesystem, or download-and-run primitives found in scanned source. PASS = no static red flag. Static scanning is blind to code fetched or executed at runtime.

• Version behavior

  fail10% of grade

  Diff of tool definitions between versions; new permissions or changed descriptions in a patch release (the postmark-mcp class).

  vs prior 1.17.0 (patch bump): new hardcoded outbound host/BCC destination(s): api.bra.in.

• Transport config

  fail10% of grade

  Remote servers: TLS and auth mode (none/token/OAuth). Local servers: whether the manifest indicates it phones home.

  remote MCP endpoint declares NO authentication (unauthenticated remote — nginx-ui/CVE-2026-33032 class)