dynamodb-access-patterns
Pass
Audited by Gen Agent Trust Hub on Sep 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill guides a workflow where arbitrary user-provided information is formalized and stored for ingestion by other agent skills in a multi-step design pipeline.
- Ingestion points: Descriptions of the application domain, actors, entities, and specific access patterns gathered during the interactive conversation in SKILL.md.
- Boundary markers: The reference/output-template.md file defining the output structure does not utilize delimiters or specific instructions to inform downstream agents to ignore instructions potentially embedded within the user-supplied text.
- Capability inventory: The skill possesses the capability to write output to markdown files on the local filesystem.
- Sanitization: There is no explicit process for sanitizing or escaping user-provided descriptions before they are written to the final document.
- Risk Factor: A downstream agent (such as the dynamodb-table-design skill mentioned in the pipeline) could interpret malicious instructions hidden within a user's domain description as legitimate system directives.
Audit Metadata