netdata-instrumentation

Pass

Audited by Gen Agent Trust Hub on Oct 13, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill provides instructions to download instrumentation components, including the OpenTelemetry .NET auto-installation script and the Java agent, from the official OpenTelemetry organization's releases on GitHub.
  • [REMOTE_CODE_EXECUTION]: The documentation for .NET instrumentation includes a process for downloading a shell script (otel-dotnet-auto-install.sh) via curl and executing it using sh. This is a standard installation pattern for the OpenTelemetry project.
  • [COMMAND_EXECUTION]: The skill contains numerous shell commands for package management (npm, pip, gem, pecl, composer) and application lifecycle management (dotnet run, go get, python app.py) that the agent may be asked to execute on behalf of the user.
  • [INDIRECT_PROMPT_INJECTION]: The skill's verification steps instruct the agent to query a Netdata Agent's API or use MCP tools to inspect metrics. This introduces an attack surface where an attacker who can manipulate the monitoring data could potentially inject instructions into the agent's context. Ingestion points: API responses from http://NETDATA_HOST:19999/api/v2/contexts and output from metric listing tools described in SKILL.md and the language-specific rules. Boundary markers: The instructions lack explicit delimiters or warnings to treat the monitoring data as untrusted. Capability inventory: The skill facilitates significant system interactions, including environment variable configuration, package installation, and execution of application code. Sanitization: No data sanitization or filtering is specified for the ingested metric information.
  • [DYNAMIC_EXECUTION]: The instrumentation guides for Node.js, Python, and Ruby instruct users to create local initialization scripts (e.g., instrument.js, instrument.py, opentelemetry.rb) that are then loaded or imported into the main application process at startup.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 13, 2026, 12:24 PM