omnibus-instrument-logs

Pass

Audited by Gen Agent Trust Hub on Mar 28, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes package manager commands (npm install, pip install, go get, mvn) to install OpenTelemetry SDKs and exporters as part of the setup process.
  • [EXTERNAL_DOWNLOADS]: The skill downloads official OpenTelemetry libraries from standard registries (NPM, PyPI, Maven Central, Go Proxy) to enable logging functionality.
  • [DATA_EXFILTRATION]: The skill is designed to transmit application logs to PostHog's ingestion endpoint (https://us.i.posthog.com/i/v1/logs). This is the intended primary purpose of the skill and uses standard OTLP headers for authentication.
  • [CREDENTIALS_UNSAFE]: The skill explicitly instructs the agent to use environment variables for PostHog API keys and OTLP endpoints, leveraging MCP tools (check_env_keys, set_env_values) rather than hardcoding secrets in source code.
  • [INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it ingests and analyzes untrusted data from the local codebase and pull requests to determine logging requirements.
  • Ingestion points: Reads files from the codebase to detect platforms and logging libraries; analyzes PR content (SKILL.md).
  • Boundary markers: None specified in instructions to distinguish between code and potential embedded instructions.
  • Capability inventory: Perform file writes, install packages via shell, and set environment variables.
  • Sanitization: No explicit sanitization or validation of the codebase content before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 28, 2026, 07:13 AM