knowledge-priming-refiner
Pass
Audited by Gen Agent Trust Hub on May 26, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes untrusted data from repository manifests which creates an indirect prompt injection attack surface.
- Ingestion points: Scans
package.json,Cargo.toml,go.mod,pyproject.toml, and.lattice/config.yamlto inform the agent's behavior during the interview. - Boundary markers: Absent; there are no explicit instructions or delimiters used to prevent the agent from following instructions potentially embedded in the repository metadata files.
- Capability inventory: The skill is capable of reading project files and writing both the knowledge-base document and configuration updates to
.lattice/config.yaml. - Sanitization: No validation or sanitization of the content from the project manifests is performed before it is used in the conversation context.
Audit Metadata