leadmagic-bulk-enrichment
Pass
Audited by Gen Agent Trust Hub on Oct 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external CSV data for enrichment and CRM export, which represents an indirect prompt injection surface. However, the skill incorporates multiple quality gates, intake validation rules (Phase 1), and QA sampling (Phase 6) to mitigate risks associated with untrusted data.
- Ingestion points: CSV input files described in
SKILL.mdandreferences/batch-pipeline-spec.md. - Boundary markers: The process includes an Intake QA stage and specific rules for rejecting rows based on domain and email criteria.
- Capability inventory: The workflow involves data enrichment, CRM upserts, and sequencer exports.
- Sanitization: Includes status-based routing and a manual QA sample step before full export.
- [COMMAND_EXECUTION]: Includes a Python script
scripts/check-output.pyused for local validation of deliverables. The script performs regex-based checks on local files to ensure they meet the defined standards.
Audit Metadata