One node of the ThetaGrip brain · HouseRule
VOICE DICTATION AUTO-CORRECT (compensate for voice-to-text, by fixed table)
HouseRule:docs/architecture/claude-rules/doors-and-surfaces.md#voice-dictation-auto-correct-compensate-for-voice-to-text-by-fixed-table
Where it lives
docs/architecture/claude-rules/doors-and-surfaces.md:41
The line
Voice-to-text reliably mangles the load-bearing nouns (qwen→"Quinn", SQLite→"sequel light", repo→"Rico", paradox voice→"paradox boys", bifurcate→"by furcate") and Claude then burns budget guessing the term instead of…
The query
This node with its neighbours, one hop out. Copy it; it runs when the view opens.
MATCH (n:Node {id: 'HouseRule:docs/architecture/claude-rules/doors-and-surfaces.md#voice-dictation-auto-correct-compensate-for-voice-to-text-by-fixed-table'})-[r]-(m) RETURN n, r, m LIMIT 50What it touches · 2
SHAPES
Request access
We built this map for our own brand. We build the same map for your company.
The install and the measurement are free and open-source (MIT); only the capability (gauntlet, stakeholder stacks, map of maps, reefs) and underwriting are licensed.
The graph is the proprietary layer: which rule shapes which sentence, placed on the 144 cells, with the file and line each one came from.
The navigating Neo4j view opens after login. It is coming soon; it is not live yet.
A request is logged with the node you were reading. A person reads the log and writes back.