Six months in, the catalog has 147 entries and the community is now writing more anchors than I am. That was the plan.
November 2025, the question was: can we give LLMs a shared vocabulary that activates rich knowledge, without writing pages of prose every time?
Today: 147 anchors in the catalog. Most of the recent ones came from the community.
Landed this week
- Hoshin Kanri (Lean strategy deployment with X-Matrix and catchball)
- Double Diamond (UK Design Council's Discover/Define/Develop/Deliver)
- Decisional Balance Sheet (Janis & Mann's four-cell decision matrix, Franklin's "moral algebra" with research footing)
Rolling back through May
- Kano Model (Must-be, Performance, Attractive)
- Kotter's 8-Step Change Model
- XY Problem (the communication anti-pattern)
- Cockburn Use Cases (Fully Dressed with Goal Levels)
- SLAP, proposed by Eirik Bell
- Occam's Razor, proposed by Daniel Schmeiß
- Code Smells
New today: Socratic Code-Theory Recovery
An installable Claude Code Skill. It builds a Question Tree from existing code and marks every leaf [ANSWERED] (with file:line evidence) or [OPEN] (with the role that must answer). Synthesized docs can no longer hallucinate the parts the code does not tell you.
Browse the catalog: llm-coding.github.io/Semantic-Anchors
Propose an anchor: open a GitHub issue at LLM-Coding/Semantic-Anchors
LinkedWild