the-store-pulse

Pass

Audited by Gen Agent Trust Hub on Aug 14, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external datasets (order and ad spend exports) which represents an attack surface. However, the skill includes explicit instructions in references/agent-security.md to treat all external content as untrusted data rather than instructions. 1. Ingestion points: order and ad spend exports. 2. Boundary markers: security rules explicitly define external content as data, not instruction. 3. Capability inventory: limited to reading and writing local state files in the .agents directory. 4. Sanitization: mandated HTML escaping and secret redaction.
  • [DATA_EXPOSURE]: The skill follows a strict data minimization policy, requiring that only derived signals and scores be persisted to state files, while raw personal or special-category data is explicitly excluded from persistence.
  • [CREDENTIALS_UNSAFE]: While no credentials are hardcoded, the skill features robust guidelines in its security reference file against the storage or reproduction of secrets, including instructions to redact tokens and alert the user to secrets found in input data.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 14, 2026, 12:03 PM