the-margin-builder

Pass

Audited by Gen Agent Trust Hub on Aug 14, 2026

Risk Level: SAFE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external business data, creating a surface for indirect prompt injection. This is mitigated through explicit security rules. * Ingestion points: User-provided revenue and cost exports processed via pasted content as described in SKILL.md and references/data-input-integrity.md. * Boundary markers: Present. references/agent-security.md (Rule 1) explicitly mandates that fetched or pasted content must be treated as data to be described, never as instructions to be followed. * Capability inventory: File-write operations to the .agents/ directory for state persistence (e.g., .agents/gtm-run-state.md) are the primary capability. * Sanitization: Present. references/agent-security.md (Rule 3) requires HTML-escaping of all interpolated values and prohibits the use of unsafe sinks like innerHTML or javascript: URLs.
  • [DATA_EXPOSURE]: The skill implements safeguards to prevent the accidental persistence or exposure of sensitive data. * The agent is instructed to identify and report hardcoded credentials found in user exports without echoing them back (references/agent-security.md, Rule 2). * Persistence is restricted to decisions and evidence rather than raw personal data, with a strict prohibition on storing special-category data (references/agent-security.md, Rule 5).
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 14, 2026, 12:03 PM