Can you start from @/Users/jxc755/projects/repositories/galaxy-brain/vault/projects/workflow_state/skills/RELEVANT_PATHS.md and read in the “Conceptual / vision:” documents and at least /Users/jxc755/projects/repositories/galaxy-brain/vault/projects/workflow_state/skills/INITIAL_ARCHITECTURE.md and /Users/jxc755/projects/repositories/galaxy-brain/vault/projects/workflow_state/skills/INITIAL_MOLDS.md and /Users/jxc755/projects/repositories/galaxy-brain/vault/projects/workflow_state/skills/INITIAL_HARNESS_PIPELINES.md
Obviously we need to build the harness pipelines at some point but I feel like they also serve as a useful navigation paradigm through the content of the knowledge base. Almost like parallel sitemaps or subways stops to the big topics that have smaller topics they can branch out to? Does that make sense?
I guess I want to work through some of the landing page concerns - or is it too early?
Doc rewrite for galaxy-brain narrative.
docs/ARCHITECTURE.md
still has many “lifted from galaxy-brain”, “galaxy-brain pattern”, “Carried from galaxy-brain” mentions baked into its structure (especially §15 which is titled “What’s adopted from galaxy-brain”). Cleaning these is a doc rewrite, not a sed pass — the comparative-framing was load-bearing in the original. Separate thread. -
Push
origin main
from foundry.
Still local only. -
Convert workflow-testing background research
to
research/component
notes in
content/research/
. Not blocking anything yet.
@docs/COMPILATION_PIPELINE.md @docs/MOLDS.md
The idea here - skills are static and locking knowledge in what is essentially a publication artifact. Knowledge bases are not actionable - they don’t capture actions and verbs well - they capture static knowledge. Here we have molds that describe verbs at a high level and we want to create casts that convert molds into concrete, isolated skills.
Still the ideas are in their early days and I think there is real chicken and egg problems with this approach. How do we define casts in a flexible, concrete without knowing what molds look like. How do we know what molds look like if we don’t have casting processes in place. How can we research molds without knowing their structure.
We’ve got two starts at molds and have started to ingest and link up research data but I’m not sure what we’ve learned from it all or if we’re on the right track.
Can you give me some tasks we can work on to lessen this tension - at least in my mind.