LSP/Index Engineer
Pass
Audited by Gen Agent Trust Hub on Oct 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill uses shell commands to install and interact with language servers.
- Evidence:
npm install -g typescript-language-server,intelephense,gopls,rust-analyzer,pyright. - Evidence: Executes language servers as subprocesses using standard
--stdioflags for JSON-RPC communication. - [EXTERNAL_DOWNLOADS]: The instructions direct the agent to download and install common developer tools and language servers from official registries.
- Evidence: Installation of well-known packages like
typescript,pyright, andgoplsvia standard package managers. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data (project source code) to build semantic graphs, which creates a potential surface for indirect injection if malicious instructions are embedded in code comments or symbols.
- Ingestion points: Reads local project files (
**/*.{ts,tsx,js,jsx,php}) and extracts content via LSP requests (textDocument/definition,textDocument/hover). - Boundary markers: None specified in the logic for separating data from instructions during processing.
- Capability inventory: Subprocess execution (LSP binaries), file system reading, and network operations (HTTP/WebSocket server).
- Sanitization: No explicit sanitization or validation of the content retrieved from language servers is mentioned before it is processed into the graph nodes or navigation index.
Audit Metadata