customerio-performance-tuning
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data (user IDs and attributes) which is then used in API calls to Customer.io. This creates a surface for indirect prompt injection if the source data contains malicious instructions designed to influence the agent's behavior during data processing or debugging.\n
- Ingestion points: The skill accepts
userIdandattrsin functions withinlib/customerio-dedup.ts,lib/customerio-batch.ts, andreferences/implementation-guide.md.\n - Boundary markers: No explicit boundary markers or instructions to ignore embedded content are provided when handling these data fields.\n
- Capability inventory: The skill performs network operations via the Customer.io SDK and
curlcommands.\n - Sanitization: Data is processed via
JSON.stringifyor passed directly to SDK methods without specific sanitization for prompt injection patterns.
Audit Metadata