the-lifecycle-mapper

Pass

Audited by Gen Agent Trust Hub on Aug 14, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external data sources such as CRM exports, call transcripts, and inbound replies, which constitutes an indirect prompt injection surface.
  • Ingestion points: The skill reads and processes content from fetched pages, CSV/pasted exports, and customer communication transcripts.
  • Boundary markers: The skill includes explicit instructions in references/agent-security.md to treat all retrieved content as data rather than instructions, and it provides a specific protocol for quoting injection attempts rather than summarizing them to prevent propagation.
  • Capability inventory: The skill reads configuration from .agents/product-context.md and writes state snapshots to .agents/gtm-run-state.md.
  • Sanitization: The skill mandates HTML-escaping for all interpolated values in emitted markup and requires that all URLs be validated to https: or relative paths before use.
  • [DATA_EXPOSURE]: The skill manages customer lifecycle data but implements strong data minimization and privacy safeguards. It explicitly forbids the persistence of special-category data (e.g., health, religion, political affiliation) and requires the agent to inform the user of every file path where customer data is being written.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 14, 2026, 12:03 PM