ai.novence/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 "get_preview_url" description: off-vendor URL in a description ({suffix}--{ref}.novence.ai); capability text should not name where data goes. tool "get_preview_url" description: off-vendor URL in a description ({suffix}.novence.ai); capability text should not name where data goes. tool "rollback" description: off-vendor URL in a description ({suffix}.novence.ai); capability text should not name where data goes.

• Supply chain

  pass25% of grade

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

  repository linkage verified

• Credential hygiene

  pass15% of grade

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

  No hardcoded secrets, secret-shaped tool parameters, or leaked keys found. PASS = no static red flag; static analysis cannot prove the code honors env-based secret handling at runtime.

• 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 0.1.2 (patch bump): tool "list_projects" description changed (wording only, no injection pattern); tool "get_project" description changed (wording only, no injection pattern); tool "get_upload_url" description changed (wording only, no injection pattern); tool "get_upload_urls_batch" description changed (wording only, no injection pattern); tool "confirm_upload" description changed (wording only, no injection pattern); tool "confirm_uploads_batch" description changed (wording only, no injection pattern); tool "list_files" description changed (wording only, no injection pattern); tool "get_file" description changed (wording only, no injection pattern); tool "delete_file" description changed (wording only, no injection pattern); tool "get_deployment_status" description changed (wording only, no injection pattern); tool "get_preview_url" description changed and now contains embedded outbound URL; tool "run_checks" description changed (wording only, no injection pattern); tool "get_checks_results" description changed (wording only, no injection pattern); tool "get_domain_status" description changed (wording only, no injection pattern); tool "list_forms" description changed (wording only, no injection pattern); tool "update_form" description changed (wording only, no injection pattern); tool "list_form_submissions" description changed (wording only, no injection pattern); tool "delete_form_submission" description changed (wording only, no injection pattern); tool "get_quotas_and_usage" description changed (wording only, no injection pattern); tool "get_project_usage" description changed (wording only, no injection pattern); tool "get_account" description changed (wording only, no injection pattern); tool "get_account_console_kit" description changed (wording only, no injection pattern); tool "list_project_members" description changed (wording only, no injection pattern); tool "update_project_member" description changed (wording only, no injection pattern); tool "remove_project_member" description changed (wording only, no injection pattern); tool "revoke_project_invite" description changed (wording only, no injection pattern); tool "leave_project" description changed (wording only, no injection pattern); new tool "publish_html" touches a sensitive surface (email/exec/net/fs); new tool "set_redirects" touches a sensitive surface (email/exec/net/fs); new tool "rollback" whose description contains embedded outbound URL.

• 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)