telnyx-texml-go

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.70). The SKILL.md shows the skill fetching and returning user-generated third-party content — for example GET /texml/Accounts/{account_sid}/Transcriptions/{...}.json (which returns transcription_text) and initiating calls where Telnyx "will request TeXML from the XML Request URL" (voice_url/Url) — meaning the agent ingests untrusted external TeXML/recording/transcription content that can materially influence subsequent actions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill instructs Telnyx to fetch TeXML from an external XML Request / voice_url at runtime (example shown as https://example.com), and that fetched TeXML directly controls call prompts/instructions which the skill requires to operate.