outbound-call-skill-creator
Pass
Audited by Gen Agent Trust Hub on Aug 4, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface as it ingests untrusted data from external sources to generate 'call goals' for telephony agents. 1. Ingestion points: Records are sourced from Google Forms, TikTok Ads, Notion, Airtable, and CSV files via MCP or local access. 2. Boundary markers: The skill instructs the agent to use 'runtime gates' and redact summaries. 3. Capability inventory: The generated workflows utilize MCP tool calls for call management and Node.js for validation. 4. Sanitization: Call goals are compiled from specific sampled fields rather than raw user-supplied instructions.
- [COMMAND_EXECUTION]: The skill provides a local Node.js script (
scripts/check-generated-skill.mjs) to ensure generated skills comply with structural requirements. The script uses string concatenation within regular expressions (e.g., inUNSUPPORTED_BINDING_LEVEL_RE) to define its validation patterns.
Audit Metadata