ads-north-star-strategy
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process potentially untrusted external data, such as brand facts from research, campaign briefs from intake, and product catalog information, to formulate advertising strategies and creative concepts. This creates a standard surface for indirect prompt injection where instructions embedded in the input data could influence agent behavior.
- Ingestion points: Data enters the agent context through the brand file intake, campaign briefs, and product research tools as described in SKILL.md and references/gooseworks-adapter.md.
- Boundary markers: The instructions do not define the use of boundary markers, such as XML delimiters or explicit 'ignore embedded instructions' warnings, when interpolating this external data into the generation process.
- Capability inventory: The skill possesses the capability to write documents (strategy.md, state.md), update campaign database records (campaign_upsert), and initiate requests for creative generation (request_campaign_generation).
- Sanitization: No explicit logic for the sanitization, escaping, or validation of business data is specified in the skill's workflow or decision rules.
Audit Metadata