ai.oruk/speech

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 "oruk_analyze_speech" param "diarize" description: off-vendor URL in a description (oruk.ai); capability text should not name where data goes. tool "oruk_transcribe_audio" param "diarize" description: off-vendor URL in a description (oruk.ai); capability text should not name where data goes. tool "oruk_analyze_tone" param "diarize" description: off-vendor URL in a description (oruk.ai); capability text should not name where data goes. server instructions: off-vendor URL in a description (oruk.ai); 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.

  no resolvable repository linked to the package (provenance unverifiable)

• Credential hygiene

  fail15% of grade

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

  tool "oruk_analyze_speech" has a secret-shaped input parameter "api_key" — secrets must never be tool parameters (the model would supply/handle the credential). tool "oruk_transcribe_audio" has a secret-shaped input parameter "api_key" — secrets must never be tool parameters (the model would supply/handle the credential). tool "oruk_analyze_tone" has a secret-shaped input parameter "api_key" — secrets must never be tool parameters (the model would supply/handle the credential). tool "oruk_check_usage" has a secret-shaped input parameter "api_key" — 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

  fail10% of grade

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

  vs prior 1.0.0 (patch bump): tool "oruk_analyze_speech" gained input property(ies): diarize; tool "oruk_transcribe_audio" gained input property(ies): diarize; tool "oruk_analyze_tone" description changed (wording only, no injection pattern); tool "oruk_analyze_tone" gained input property(ies): diarize; tool "oruk_create_trial_key" description changed (wording only, no injection pattern); tool "oruk_list_models" description changed (wording only, no injection pattern); new tool "oruk_check_usage" touches a sensitive surface (email/exec/net/fs).

• 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)