neovim-plugin-development
Pass
Audited by Gen Agent Trust Hub on Sep 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes text from Neovim buffers to perform operations like JSON encoding or formatting.
- Ingestion points: Data is read from the active buffer in
assets/plugin-template/lua/example/init.luaand referenced guidelines. - Boundary markers: The instructions do not specify prompt boundary markers for buffer content, though the example uses structural transformation (JSON encoding) which limits instruction leakage.
- Capability inventory: The skill demonstrates use of Neovim APIs (
vim.api) for buffer modification andvim.systemfor external process execution (e.g., formatters). - Sanitization: The template utilizes the built-in
vim.json.encodefor data processing. The skill guidelines strongly advocate for using host-provided APIs for serialization and process execution to ensure robustness and security.
Audit Metadata