kingscript-code-generator
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process local markdown documentation, structured JSON configuration, and Java Archive (JAR) metadata. This ingestion surface could potentially be used for indirect prompt injection if external or untrusted documentation files containing malicious instructions are introduced into the search path.
- Ingestion points: Local documentation files in the
references/directory,local-paths.json, andAGENTS.md(specified inSKILL.md). - Boundary markers: The instructions provide structure on how to parse resources but lack explicit delimiters or "ignore embedded instructions" warnings for the content of the documentation files.
- Capability inventory: The skill relies on the agent's ability to execute local shell commands such as
rg(ripgrep), PowerShellGet-ChildItem, andjarfor searching and inspecting resources (found inSKILL.md). - Sanitization: Instructions require the agent to verify that paths exist before use but do not specify sanitization of the content read from files.
- [COMMAND_EXECUTION]: The skill instructs the agent to use standard local search utilities (ripgrep, PowerShell, and JAR) to find relevant code snippets and SDK definitions. These commands are localized to the user-configured resource paths and are consistent with the skill's purpose as a code generator and technical expert.
Audit Metadata