2bbb1871cc
Architecture restructure from morning review: 1. code/scripts/ subdivided into sim/, reach/, barrier/, plot/. Easier nav; `barrier/` is the natural place for SOS scale-up scripts. 2. Heatup PJ reach variants consolidated behind TOML configs. reach_heatup_pj.jl now takes `--config path/to/config.toml`; configs/heatup/baseline.toml (wide entry, from predicates.json) and configs/heatup/tight.toml (narrow entry, reproduces all-6-halfspaces discharged result). Old reach_heatup_pj_tight.jl and reach_heatup_pj_tight_full.jl deleted (superseded). 3. Reach output .mat files moved from reachability/ to results/. reachability/ now = specs + docs; results/ = ephemeral outputs (gitignored *.mat). README added. 4. OVERNIGHT_NOTES.md archived to claude_memory/2026-04-20-21-overnight- session-summary.md (date range in the filename makes the history clearer). All include() / Pkg.activate() paths in scripts updated for the new depth. Smoke tests pass (reach_operation.jl generates its .mat in the new results/ location; sim_sanity.jl matches MATLAB). Presentation outline for the 20-min prelim talk landed in presentations/prelim-presentation/outline.md. 14-slide assertion- evidence format targeting OT-informed cybersecurity audience. Each slide: one declarative assertion + one figure. Outline includes which figures already exist and which need to be created, timing checkpoints, cybersecurity angle to emphasize, and Q&A prep. New config configs/heatup/with_steam_dump.toml + its companion scripts/reach/reach_heatup_pj_sd.jl (12-state RHS with Q_sg as an augmented bounded parameter x[10] and time as x[11]). Kicks off point 3 from morning review. Next up: scram X_entry expansion (morning point 2) — LOCA scenario + union of mode reach envelopes. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
claude_memory/
Session notes and scratch knowledge that hasn't (yet) earned a place in a
CLAUDE.md. Anything that leaves a context window uncaptured is lost
forever — this folder exists so session-level insight doesn't evaporate.
Purpose
A Claude context window is finite. Anything an agent figures out in a session
that isn't written down is gone when the session ends. The curated
CLAUDE.md files hold stable, authoritative knowledge; this folder holds
everything else that's still worth remembering.
Pattern
- One file per session with non-trivial insight.
- Filename:
YYYY-MM-DD-short-topic.md. Date first solsis chronological. - Freeform. Rough edges encouraged. Don't polish — capture.
- At minimum, answer: what did we do, what did we decide, what surprised us, what's still open.
- Cross-reference files you edited, alternatives you considered, dead-ends you hit.
What graduates vs. what stays
Three futures for a note in here:
- Graduates up into a
CLAUDE.md(ordocs/architecture.md, or a code comment) when the knowledge stabilizes and becomes authoritative. After graduating, consider deleting the note or leaving it as historical context. - Stays here as historical context — useful for "why did we do X last month?" or "what have we already tried?"
- Gets deleted when it's no longer relevant (a fixed bug, an abandoned approach, a stale investigation).
The graduation step is a discipline, not automation. When you're finishing a
session and think "a future agent needs this to avoid re-discovering it,"
ask: is this stable enough for a CLAUDE.md, or is it still forming?