Doctor Bones carries enough brain matter to reason with architecture patterns.
A Doctor Bones-based repository is not just a place to store code. It is a place to store project cognition: what the work means, what boundaries matter, which examples to read, what checks prove progress, and which architecture patterns can guide the next move.
What pattern fit means here
Doctor Bones can carry lightweight embedded reference lenses. They do not replace full source repositories, but they give a foreground AI enough structure to do a useful first-pass comparison when the human asks a simple question.
Go analyze my other repo <repo URL> with PFEM-lite analysis capabilities.
Go analyze my other repo <repo URL> with PFCOMM-lite analysis capabilities.
Use both lenses to explain this project's evidence-to-coordination boundary.
Two useful lenses
PFEM-lite
Evidence governance
PFEM-lite helps the AI ask what counts as evidence, where it came from, how confidence and provenance are preserved, and whether reports or summaries have a durable proof path.
PFCOMM-lite helps the AI ask what is being requested, who has authority, what was assigned, what status came back, and how decisions and after-action accountability are recorded.
intent
authority context
tasking
assignment
resource
status
receipt
decision log
after-action record
Why it matters
Plain prompt
The human asks normally
The human should not have to hand-write every architectural checklist. A good repo lets the human ask for a goal in normal language.
Repo cognition
The repo narrows the lane
AGENTS.md, examples, workorders, checks, and embedded reference lenses tell the foreground AI what good work means here.
Bounded execution
The executor gets a contract
If files need to change, the foreground AI creates a bounded workorder with scope, constraints, checks, and completion rules.
A Miller-flavored reading
James Grier Miller's living-systems vocabulary is useful historical inspiration for thinking about systems, flows, memory, boundaries, and adaptation. Doctor Bones does not need to become a Miller project to borrow the better instinct: a working system should preserve its own memory, signal paths, and adaptation rules.
In Doctor Bones terms, the repository should help future AI sessions recognize the project organism before operating on it.
Boundary: PFEM-lite and PFCOMM-lite are lightweight embedded references. They are enough for first-pass pattern analysis, not a substitute for inspecting the full live reference repositories when the human asks for a full comparison. External repositories stay read-only unless the human explicitly authorizes otherwise.