dd-instrument-rum
Pass
Audited by Gen Agent Trust Hub on Aug 27, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands for package installation (e.g.,
npm install,yarn add,pnpm add,bun add) and project builds (e.g.,npm run build,ng build,next build) to verify that instrumentation does not break the application. - [INDIRECT_PROMPT_INJECTION]: The skill analyzes untrusted project files, including manifests (
package.json), lockfiles, and source code, to determine the framework and required configuration. This represents a vulnerability surface where a malicious project could attempt to influence the agent's behavior. - Ingestion points: Project root manifests, lockfiles (e.g.,
pnpm-lock.yaml), source entrypoints, and framework-specific configuration files (e.g.,tsconfig.json). - Capability inventory: The agent can read/write files, install packages, and execute arbitrary build scripts defined in the project manifest.
- Boundary markers: The skill provides strict 'Ground rules' requiring the agent to inspect files before editing and to copy symbols exactly from provided reference documentation rather than relying on memory or substituted values.
- Sanitization: The instructions emphasize verifying application behavior through a terminating build, but do not specify internal sanitization of data read from the project files.
- [EXTERNAL_DOWNLOADS]: The skill fetches and installs official Datadog Browser RUM SDK packages from npm/yarn/pnpm/bun and references official Datadog CDN assets (e.g.,
https://www.datadoghq-browser-agent.com/). These resources originate from a well-known service provider and are required for the skill's stated purpose.
Audit Metadata