product-tracking-implement-tracking
Pass
Audited by Gen Agent Trust Hub on Oct 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from configuration files to generate executable code. A malicious tracking plan could potentially include instructions designed to influence the agent's output.
- Ingestion points: Reads files including
.telemetry/tracking-plan.yaml,.telemetry/delta.md,.telemetry/instrument.md, and.telemetry/product.md. - Boundary markers: The skill lacks explicit separators or instructions to ignore embedded prompts within the plan content.
- Capability inventory: The agent is instructed to write and modify multiple files in the local
tracking/directory. - Sanitization: There is no specified logic for sanitizing or validating the contents of the ingested YAML or Markdown files before they are used to generate code structures.
- [EXTERNAL_DOWNLOADS]: The skill provides instructions for integrating well-known analytics SDKs and scripts from providers such as Segment, Amplitude, Mixpanel, and PostHog. It also references the author's own infrastructure (accoil.com) for tracker scripts and API endpoints.
- Source details: References official CDNs and package registries (e.g., cdn.accoil.com, cdn.mxpnl.com, npmjs.org) for telemetry instrumentation.
Audit Metadata