@ruvector/graph-wasm

Pass

Audited by Gen Agent Trust Hub on Mar 5, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill documentation refers to the installation of packages from the npm registry using npx.
  • Evidence: npx ruvector@latest and npx @ruvector/graph-wasm@latest are suggested in the installation section.
  • [EXTERNAL_DOWNLOADS]: The initialization function allows for fetching WebAssembly binaries from external sources.
  • Evidence: await init(wasmUrl?: string | URL) allows specifying a remote path for the WASM module.
  • [COMMAND_EXECUTION]: The skill documentation describes command-line usage for installing the library environment via npx.
  • Evidence: Installation commands include npx ruvector@latest.
  • [REMOTE_CODE_EXECUTION]: The skill utilizes WebAssembly, which involves loading and executing compiled code in the runtime environment.
  • Evidence: The init() function loads and executes the @ruvector/graph-wasm WASM module.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection if untrusted data is included in Cypher queries or vertex properties without sanitization.
  • Ingestion points: Data enters the context via gdb.addVertex(), gdb.addEdge(), and the params argument in gdb.query().
  • Boundary markers: No explicit boundary markers or 'ignore' instructions are provided in the documentation templates.
  • Capability inventory: The skill can execute complex Cypher queries and serialize/deserialize graph data, though it lacks direct file system or network access (outside of WASM loading).
  • Sanitization: No specific sanitization or escaping guidelines for Cypher query parameters are mentioned in the documentation.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 5, 2026, 05:22 PM
Security Audit — agent-trust-hub — @ruvector/graph-wasm