rstore-monospace
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCREDENTIALS_UNSAFEEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill describes functionality for loading and processing external OpenAPI documents and system schema metadata (such as meta collections and constraints) from remote projects or local JSON files. This creates a potential surface for indirect prompt injection if the processed data contains malicious instructions targeting the agent interpreting these schemas.
- Ingestion points:
loadMonospaceCollectionsacceptsurl,project,input(OpenAPI JSON), andmetadataInput(metadata snapshot). - Boundary markers: The instructions do not specify any delimiters or safety warnings to ignore instructions embedded within the processed schemas.
- Capability inventory: The skill describes capabilities for network operations via a REST client, file reading for local schemas, and data mutation (CRUD) through rstore collection APIs.
- Sanitization: No sanitization or validation logic for the ingested external schema data is described in the workflow.
- [CREDENTIALS_UNSAFE]: The skill references several types of secrets including
apiKey,schemaApiKey, andruntimeApiKey. These are documented as configuration parameters for APIs likecreateMonospaceRestClientandloadMonospaceCollections. The documentation correctly advises maintainingschemaApiKeyon the build/server side, which is a standard security practice. No hardcoded credentials or private keys were detected. - [EXTERNAL_DOWNLOADS]: The documentation references several Node.js packages such as
@rstore/monospace,rstore-vue,rstore-vite-monospace, andrstore-nuxt-monospace. These resources are consistent with the skill's purpose and appear to be official components belonging to the vendor ecosystem.
Audit Metadata