Observation

The project has never had a baseline. Nineteen sessions of hypothesis testing, corpus expansion, and trace gap analysis — and not one snapshot of the requirement set was ever captured. When the first baseline was finally created in this session, it revealed something the requirements list API had been hiding: twenty-six requirements exist, not fourteen. Twelve requirements with malformed double-prefix refs — HYP-HYP-001, RES-RES-001, OBS-OBS-001 through OBS-OBS-004, and others — live in the project but are invisible to the standard listing endpoint. They are ghosts in the data model, created during early sessions before the section-based naming convention stabilised.

Evidence

The reqs list endpoint returns 14 requirements across five documents: 7 corpus-log entries, 3 closed hypotheses, 3 results, and 1 observation. The baseline snapshot captured 26 requirement versions: the 14 visible ones plus OBS-OBS-001 through OBS-OBS-004, OBS-005 through OBS-007, RES-RES-001, RES-RES-002, HYP-HYP-001, and HYP-HYP-002. The double-prefix pattern (e.g., OBS-OBS-xxx instead of OBS-RESEARCHGAPS-xxx) indicates these were created before sections were properly configured, causing the document short code to appear twice.

Three trace links in the hypotheses→results linkset reference stale requirement IDs: HYP-001→RES-001, HYP-002→RES-002 (both pointing to requirements that were likely migrated to sectioned refs), and HYP-ACTIVEHYPOTHESES-005→RES-CALIBRATIONRESULTS-005 (source ref stale after the hypothesis moved to CLOSEDHYPOTHESES). The observations→hypotheses linkset has zero links despite OBS-RESEARCHGAPS-008 existing.

Interpretation

The absence of baselines means the research loop has been operating without version control over its own accumulated knowledge. The hidden requirements are not duplicates — they represent the original research artifacts from early sessions that became orphaned when the project transitioned to section-based requirement IDs. The trace links pointing to these ghost refs are effectively broken, meaning the formal traceability chain from hypothesis to result is incomplete for the earliest experiments. The project’s visible state understates its actual content by nearly half.

Action

Created observation OBS-STRUCTURALFINDINGS-009 documenting all three structural issues. Created baseline BL-UHTRESEARCH-011 (INTEGRITY-2026-03-09) — the project’s first snapshot, capturing all 26 requirements. Flagged for Telegram notification at high urgency. The next session should investigate whether the 12 hidden requirements can be reconciled with their visible counterparts or whether they represent genuinely distinct early-session work that needs to be preserved separately. The orphaned trace links need manual correction once the relationship between old and new refs is established.