diagnosing-failed-warehouse-syncs
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted data from external sources, which could be exploited to influence the agent's actions.
- Ingestion points: The agent reads the
latest_errorfield from tool outputs such asexternal-data-sources-retrieveandexternal-data-schemas-list. This field contains error messages generated by external databases and services. - Boundary markers: The instructions do not specify the use of delimiters or specific prompt engineering to isolate the untrusted error strings from the agent's instructions.
- Capability inventory: The skill has access to several impactful tools, including
external-data-schemas-resync(wipes data),external-data-schemas-delete-data(drops tables), andexternal-data-sources-partial-update(modifies credentials). - Sanitization: The skill mitigates risks by requiring explicit human confirmation before executing any destructive recovery actions, as outlined in 'Step 5 — Confirm before destructive actions'.
Audit Metadata