sales-follow-up
Pass
Audited by Gen Agent Trust Hub on Sep 10, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill provides an installation command that fetches content from a GitHub repository (
github.com/peter-tu-zynkr/zynkr-skill-builder). This repository is owned by the vendor associated with the skill author. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external, potentially untrusted data in the form of sales demo transcripts provided via Google Docs or direct text input.
- Ingestion points: Transcript parsing logic described in the workflow section of
SKILL.md. - Boundary markers: The instructions do not define specific delimiters to separate transcript content from the agent's internal logic.
- Capability inventory: The skill has permissions to draft Gmail messages, create CRM notes, update deal summaries, and manage CRM tasks using specific MCP tools.
- Sanitization: The risk is mitigated by the skill's design, which maps extracted pain points against a grounded vendor solution page (
zynkr.ai/ai-platform) rather than allowing free-form execution based on transcript contents. Additionally, the Gmail output is created as a draft for mandatory human review. - [COMMAND_EXECUTION]: The documentation includes a setup command (
npx skills add ...) intended for the user to execute to install the skill. This command points to the vendor's official skill-building repository.
Audit Metadata