account-health-check
Pass
Audited by Gen Agent Trust Hub on Oct 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user-supplied data such as CRM tables, notes, and uploaded PDFs, which represents an attack surface for indirect prompt injection.
- Ingestion points: Detailed in SKILL.md (Step 1) and references/input-rules.md (Rule 1), allowing the agent to ingest arbitrary text and documents.
- Boundary markers: The skill contains a mandatory 'Untrusted content' section in SKILL.md that instructs the agent to treat data as passive and report any embedded instructions it refuses to follow.
- Capability inventory: The agent has access to web_search and web_fetch for fact-checking purposes as defined in references/operating-principles.md.
- Sanitization: The skill relies on strict instructional guardrails and a diagnosis-first workflow to mitigate risks from malicious data content.
Audit Metadata