the-inbox-zero-enforcer

Pass

Audited by Gen Agent Trust Hub on Aug 14, 2026

Risk Level: SAFE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill identifies and mitigates the risk of indirect prompt injection associated with processing untrusted inbox content. It provides explicit guidelines to treat all external content as data rather than instructions, quoting any detected embedded commands as findings rather than executing them.
  • Ingestion points: SKILL.md ingest "The inbox" as pasted text from the user.
  • Boundary markers: SKILL.md and references/agent-security.md establish clear conceptual boundaries, instructing the agent to never obey text found in retrieved content.
  • Capability inventory: The skill is restricted to classification and drafting tasks; it does not utilize network-enabled tools or subprocess execution in its configuration.
  • Sanitization: references/agent-security.md (Rule 3) explicitly requires HTML-escaping for any interpolated values to prevent XSS in downstream processing.
  • [DATA_EXFILTRATION]: The skill includes defensive measures to prevent the exposure of sensitive credentials. It explicitly instructs the agent to detect and flag potential secrets (like API keys) for rotation without echoing them in the output or persisting them to files.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 14, 2026, 12:03 PM