instrument-feature-flags
Pass
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill analyzes local codebase files to detect platforms and instrument code. This ingestion of untrusted data represents a standard surface for indirect prompt injection.\n
- Ingestion points: Dependency files (e.g., package.json, pubspec.yaml, requirements.txt), environment files (.env), and source code files as specified in Step 1 and Step 5 of SKILL.md.\n
- Capability inventory: Reading and writing to the file system (Step 5/6), network requests to PostHog APIs (Step 6), and package installation commands (documented in Reference files).\n
- Boundary markers: No explicit delimiters or warnings to ignore embedded instructions are used when processing external code content.\n
- Sanitization: No explicit filtering or sanitization of ingested code content is mentioned.\n- [EXTERNAL_DOWNLOADS]: The skill provides instructions to download and install official PostHog SDKs from standard package registries (NPM, PyPI, Maven, Crates.io, etc.) and references official vendor tools such as the PostHog wizard (npx @posthog/wizard). These resources originate from official vendor repositories and domains.
Audit Metadata