Setting the file. One moment.
Skill 07 · Auditing Experiments Flags
Subchapter 7.2
references/finding-taxonomy.mdMarkdown3 KBView on GitHub
| Severity | Badge | Meaning |
|---|---|---|
| CRITICAL | 🔴 | Blocks correctness — experiment results may be invalid or flag behavior is broken |
| WARNING | 🟡 | Needs attention — not broken yet but risks exist or best practices are violated |
| INFO | 🔵 | Suggestion — hygiene improvement, safe to defer |
| Category | Description | Max severity |
|---|---|---|
| Correctness | Integrity issues that affect experiment results or flag evaluation | CRITICAL |
| Waste | Active resources not serving a purpose (running experiments going nowhere, flags nobody evaluates) | WARNING |
| Process | Methodology and practice issues (missing hypothesis, no metrics) | WARNING |
| Complexity | Fragility and maintainability concerns (too many toggles, high churn) | WARNING |
| Cleanup | Hygiene items — stale drafts, orphaned flags, safe to defer | INFO |
| Security | PII or access concerns in flag/experiment configuration | WARNING |
Never assign a severity higher than the category’s max:
When reporting a finding, always include:
When listing multiple findings: