RISK-026-001 |
Planned cases are mistaken for completed evidence. |
high |
READY/PLANNED state and runner exclusion |
026-G03 |
RISK-026-002 |
A V&V refactor changes numerical behavior or public API. |
high |
baseline fingerprints and focused API smoke |
026-G00/026-G14 |
RISK-026-003 |
External tool absence is represented as a pass. |
high |
explicit SKIPPED_EXTERNAL_UNAVAILABLE outcome |
026-G13 |
RISK-026-004 |
Large artifacts obscure source history and inflate clones. |
medium |
artifact policy and digest-first summaries |
026-G01 |
RISK-026-005 |
Performance comparisons use incompatible hardware or one noisy sample. |
medium |
hardware metadata, repeated medians and profile bands |
026-G12 |
RISK-026-006 |
Research paths are promoted by test count alone. |
high |
claim matrix and Owner-only maturity decision |
026-G15 |