instrument-error-tracking
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands to install PostHog SDKs and manage project dependencies using platform-appropriate package managers including npm, yarn, pnpm, bun, pip, go get, bundle, and composer.\n- [EXTERNAL_DOWNLOADS]: The skill initiates the download and execution of vendor-provided tools, such as
npx @posthog/wizard, and standard library packages from established registries to automate source map uploads and environment setup.\n- [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes untrusted data from the user's codebase to drive its logic.\n - Ingestion points: Project dependency manifests (e.g., package.json, requirements.txt, pubspec.yaml) and environment configuration files (.env, .env.local) are read in STEP 1 and STEP 7 to detect the platform and existing keys.\n
- Boundary markers: The instructions do not specify the use of delimiters or provide instructions to ignore potentially malicious content embedded within the analyzed project files.\n
- Capability inventory: The skill has the capability to write to source code, modify environment files, and execute shell commands via package managers, which could be exploited if malicious instructions are ingested from project files.\n
- Sanitization: No explicit validation or sanitization is required for the content read from the codebase files before it is processed by the agent.
Audit Metadata