monospace

Pass

Audited by Gen Agent Trust Hub on Oct 2, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Fetches early-access versions of the Monospace SDK and CLI from a preview infrastructure at pkg.pr.new belonging to the developer's organization.
  • [COMMAND_EXECUTION]: Uses the monospace CLI tool to perform authentication, workspace initialization, and type generation from instance schemas.
  • [DYNAMIC_EXECUTION]: Includes a local Node.js utility script (scripts/schema-view.mjs) for decoding schema manifests and generates TypeScript source code at runtime via the CLI's codegen functionality.
  • [INDIRECT_PROMPT_INJECTION]: The skill facilitates the processing of data and schema information retrieved from remote Monospace instances.
  • Ingestion points: Retrieves schema manifests through the /schema endpoint and collection data via item-scoped REST and SDK calls.
  • Boundary markers: The instructions guide the agent to use generated types and read-back verification; however, there are no explicit boundary markers or instructions to sanitize raw text data for prompt injection.
  • Capability inventory: The skill uses npx monospace sdk generate to write files to the filesystem and node scripts/schema-view.mjs to read and parse JSON manifest files.
  • Sanitization: The logic relies on JSON parsing and generated type safety; it does not include specific sanitization or filtering for untrusted natural language content embedded in data fields.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 2, 2026, 09:43 PM