helm-chart-scaffolding
Pass
Audited by Gen Agent Trust Hub on Mar 3, 2026
Risk Level: SAFEDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill includes a 'Memory Protocol' that requires the agent to perform read and write operations on internal state files located at
C:\dev\projects\agent-studio\.claude\context\memory\. Accessing these paths could lead to the exposure of the agent's persistent context or sensitive project metadata. - [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection through its core templating functionality.
- Ingestion points: The skill ingests untrusted data from user-provided
values.yamlfiles and CLI arguments. - Boundary markers: There are no explicit delimiters or protective instructions in the provided templates to prevent malicious payloads in the values from influencing the agent or downstream systems.
- Capability inventory: The skill has access to
Bash,Read,Write,Edit,Glob, andGreptools, providing broad file system and execution privileges. - Sanitization: The workflow lacks input validation or sanitization for data interpolated into Go templates.
Audit Metadata