Kismet Context Vault

2026-05-06T1440--3

2 min read

dashboard_url: https://dashboard.kismetpets.com/context/health/lint/3/ generated_at: ‘2026-05-06T14:40:20.186668+00:00’ id: 3 scope_id: null scope_type: ”

Lint Report — 2026-05-06 10:40

Scope: global

Context Lint Report — 2026-05-06 #3 (post-incident)

Score: 70/100 (unchanged from runs #2 and #3 — orphan count is 120 now, up from 117 because 3 new results were saved between runs while 3 were ingested).

Orphan results (120)

The synthesis layer is at 6 entities + 12 concepts after run #3, all with embeddings. Three new results have appeared since the last steward run:

  • voice_mining:6 “Daily Boost Toppers — Naming Validation VOC” (2026-05-05) — fresh and high-density
  • front_end_design:3 “45-Day Gut Transformation Landing — Grid Extra Info” (2026-05-01) — landing page
  • dtc_ads:32 & dtc_ads:33 Target retail hero ads (2026-04-29) — newer than last run

The remaining 116 are pre-existing orphans deferred across runs #2 and #3 (“budget exceeded” rationale). Distribution roughly unchanged at 44 positioning, 30 dtc_ads, 26 market_research, 7 direct_response_copy, 4 voice_mining, 2 front_end_design, 2 measurement, 2 orchestrator.

Stale entities (0)

The 18 entities/concepts seeded in runs #2 and #3 are still under the 30-day staleness threshold. Will start mattering after May 28.

Duplicate candidates (0)

None at threshold 0.92. Embeddings ARE generated and the detector is working — last steward run installed them. As more entities accumulate (especially same-kind competitors and ICP segments), this number should start being non-zero.

Briefs needing compaction (0)

The lone experiment:1 brief has 3 deltas, threshold for compact is 7. Next steward run that adds 4+ ingests pointing at experiment:1 will trip the threshold and /context-compact-brief will become a high-priority recommendation.

  1. Run /context-ingest-result on the 3 new orphan results first — they’re freshest and richest, especially voice_mining:6 (recent VOC for the new Daily Boost Toppers product). Newer = more market-current insight.
  2. Continue chipping at the older 116-orphan backlog.
  3. Defer compact + merge until thresholds trip.

Graph View