website-visitor-identification
Pass
Audited by Gen Agent Trust Hub on Oct 10, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill primarily contains instructional markdown content and templates for Go-To-Market (GTM) strategies. No evidence of malicious behavior, credential theft, or unauthorized network activity was found.
- [COMMAND_EXECUTION]: A Python script (
scripts/check-output.py) is included to validate the output generated by the agent. It performs basic text checks (unresolved placeholders, length, and keyword presence) on a local file path provided as a command-line argument. The script does not execute arbitrary code or interact with the network. - [INDIRECT_PROMPT_INJECTION]: The skill describes processing external signals such as UTM parameters, referrer strings, and visitor intent data. While these represent an ingestion point for untrusted data, the skill mitigates risk by prescribing manual triage (Slack alert triage), ICP filtering, and human-in-the-loop review before taking actions like CRM creation or email outreach. No automated execution of instructions found in external data is performed.
Audit Metadata