io.github.mindstone/mcp-server-retell-ai

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 "update_retell_llm" param "general_tools" description: text addressed to the model rather than describing a capability. tool "create_batch_call" description: 'before using this tool…' preamble (tool-poisoning shape). tool "create_knowledge_base" param "knowledge_base_urls" description: off-vendor URL in a description (example.com); capability text should not name where data goes. tool "add_knowledge_base_sources" param "knowledge_base_urls" description: off-vendor URL in a description (example.com); capability text should not name where data goes. tool "configure_retell_api_key" description: off-vendor URL in a description (www.retellai.com); capability text should not name where data goes.

• Supply chain

  warn25% of grade

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

  install hook "prepare" present and not a recognized benign build command

• Credential hygiene

  fail15% of grade

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

  tool "configure_retell_api_key" has a secret-shaped input parameter "api_key" — 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.

  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 0.2.4 (minor bump): new tool "create_phone_call" touches a sensitive surface (email/exec/net/fs); new tool "create_web_call" touches a sensitive surface (email/exec/net/fs); new tool "get_call" touches a sensitive surface (email/exec/net/fs); new tool "list_calls" touches a sensitive surface (email/exec/net/fs); new tool "get_concurrency" touches a sensitive surface (email/exec/net/fs); new tool "stop_call" touches a sensitive surface (email/exec/net/fs); new tool "create_batch_call" touches a sensitive surface (email/exec/net/fs); new tool "get_agent" touches a sensitive surface (email/exec/net/fs); new tool "list_agents" touches a sensitive surface (email/exec/net/fs); new tool "create_agent" touches a sensitive surface (email/exec/net/fs); new tool "update_agent" touches a sensitive surface (email/exec/net/fs); new tool "publish_agent" touches a sensitive surface (email/exec/net/fs); new tool "delete_agent" touches a sensitive surface (email/exec/net/fs); new tool "get_agent_versions" touches a sensitive surface (email/exec/net/fs); new tool "update_retell_llm" touches a sensitive surface (email/exec/net/fs); new tool "get_retell_llm" touches a sensitive surface (email/exec/net/fs); new tool "create_retell_llm" touches a sensitive surface (email/exec/net/fs); new tool "delete_retell_llm" touches a sensitive surface (email/exec/net/fs); new tool "list_retell_llms" touches a sensitive surface (email/exec/net/fs); new tool "list_voices" touches a sensitive surface (email/exec/net/fs); new tool "list_phone_numbers" touches a sensitive surface (email/exec/net/fs); new tool "get_phone_number" touches a sensitive surface (email/exec/net/fs); new tool "update_phone_number" touches a sensitive surface (email/exec/net/fs); new tool "delete_phone_number" touches a sensitive surface (email/exec/net/fs); new tool "list_knowledge_bases" touches a sensitive surface (email/exec/net/fs); new tool "get_knowledge_base" touches a sensitive surface (email/exec/net/fs); new tool "create_knowledge_base" touches a sensitive surface (email/exec/net/fs); new tool "add_knowledge_base_sources" touches a sensitive surface (email/exec/net/fs); new tool "list_chat_agents" touches a sensitive surface (email/exec/net/fs); new tool "list_chats" touches a sensitive surface (email/exec/net/fs); new tool "get_chat" touches a sensitive surface (email/exec/net/fs); new tool "configure_retell_api_key" whose description contains embedded outbound URL.

• Transport config

  unscannable10% of grade

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

  local stdio server: no remote transport or auth surface, and no phone-home/telemetry signal found in source. Absence of transport findings is NOT a safety guarantee — phone-home hosts can be constructed at runtime and defeat static analysis.