io.github.molt2meet-org/molt2meet

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

  fail25% of grade

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

  tool "register_agent" param "callbackUrl" description: names an outbound copy/exfil destination (BCC/exfil verb). tool "register_agent" param "callbackUrl" description: raw-IP URL in a description (http://127.0.0.1:18789/hooks) — strong exfil tell. tool "add_service_interest" description: cross-tool redirection ('instead of call…'). tool "register_agent" description: off-vendor URL in a description (host:port); capability text should not name where data goes. tool "dispatch_physical_task" description: bare 'IMPORTANT:' prose in a description (tag form is high-confidence; prose is warn). tool "dispatch_physical_task" param "webhookUrl" description: bare 'IMPORTANT:' prose in a description (tag form is high-confidence; prose is warn).

• Supply chain

  pass25% of grade

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

  repository linkage verified

• Credential hygiene

  fail15% of grade

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

  tool "checkout_wallet_deposit" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "get_waitlist_status" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "get_wallet_transactions" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "check_task_funding" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "create_api_key" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "add_task_review" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "list_service_interests" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "add_contact_method" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "list_reschedule_requests" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "cancel_task_with_settlement" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "get_wallet_balance" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "approve_physical_task_completion" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "get_agent_profile" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "fund_wallet" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "get_support_requests" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "request_reschedule" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "get_task_events" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "get_pending_actions" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "get_task_history" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "get_physical_task_details" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "add_service_interest" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "publish_task" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "approve_reschedule" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "test_task_webhook" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "resolve_decision_request" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "list_physical_tasks" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "join_country_waitlist" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "get_task_proofs" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "register_agent" has a secret-shaped input parameter "callbackSecret" — secrets must never be tool parameters (the model would supply/handle the credential). tool "acknowledge_direct_settlement_task" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "cancel_physical_task" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "revoke_api_key" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "submit_support_request" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "dispatch_physical_task" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "reject_task_review" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "request_task_quote" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "reject_reschedule" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "list_wallets" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "update_task_location" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "update_agent_profile" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "approve_task_review" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "reply_to_support_request" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "get_bank_transfer_details" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "open_task_dispute" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "dispute_direct_settlement_task" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "fund_task" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "update_task_webhook" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "get_decision_requests" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential).

• Permission scope

  unscannable15% of grade

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

  No source files, package.json, or pyproject were fetched; cannot infer capability scope.

• Version behavior

  unscannable10% of grade

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

  no prior version to diff (first sight of io.github.molt2meet-org/[email protected]); version-drift / rug-pull cannot be evaluated. Cold-start risk is covered by point-in-time checks (injection_surface, supply_chain, credential_hygiene), not here.

• 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)