konnect-reference-platform
Pass
Audited by Gen Agent Trust Hub on Oct 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill presents an indirect prompt injection surface because it processes untrusted external data that could contain malicious instructions designed to influence the agent's behavior during the design or deployment phase.\n
- Ingestion points: The workflow requires reading checked-in
kongctlmanifests, decK files, and OpenAPI contracts from the repository (SKILL.md, Workflow Step 1).\n - Boundary markers: There are no specific instructions for the agent to use delimiters or to ignore embedded natural language instructions within the ingested contracts or configuration files.\n
- Capability inventory: The skill directs the agent to use tools such as
kongctl-declarativeanddeck-gatewayto apply, sync, and reconcile the state of the Konnect platform based on the ingested data (SKILL.md, Handoffs).\n - Sanitization: While the skill mandates technical validation of YAML syntax and resource diffs, it does not provide mechanisms for sanitizing or filtering natural language content within the external data sources.
Audit Metadata