BFM-decomp/decomp-architect/memory-seed/MEMORY.md

Memory Index

Already provided by ProjectArchitect 2.0's seed (cited, not duplicated)

who-is-dev · project-governance-system · autonomous-within-phases · build-tasklist-after-plan-approval · capture-knowledge-before-fresh-session · clarify-misconception-before-costly-action · justify-new-tools-before-adopting · dont-block-loop-with-askuserquestion · no-commit-co-author · commit-per-task-user-pushes · plain-english-recap · session-start-rules-in-full · worklogs-reference-only · keep-ops-current · effort-discipline · effort-prompt-ultracode-on-breadth — and the repository-resident state itself (.claude-state/, registry rule H8).

The decomp seed

• The byte gate is the only claim — "matched" means byte-identical output AND a green whole-binary hash; nothing functionally equivalent counts; "banked" is written from a tool's printed line
• The matching flywheel — consult the knowledge base before every match, feed the generalizable lesson back into BOTH the cookbook and the tooling after it
• MCP reconnect after restart — after restarting the disassembler's MCP server or switching the served program, pause and ask the developer to reconnect the client; the agent cannot
• Offline tooling first — every COMPUTABLE step becomes a deterministic zero-token tool; keep deterministic recovery and search separate; ask "is this residual computable?" before reaching for a model
• A slow gate is a bug — banking, not drafting, is the bottleneck; nothing in the gate is serial; pass -j to every build; measure the gate before fearing it
• Breadth is isolated agents — the main loop's context accumulates roughly quadratically; N isolated agents cost about linearly and run in parallel; hand-by-main-loop wins only for a handful
• Route by measured difficulty — tier drafters by a cliff you measured on your own corpus; escalate sooner; the strongest model is for new wall classes, never for reviewing a corpus
• One runbook is the procedure — one document is the operating procedure for the campaign loop, read before every run, kept current in the same change; a superseded runbook gets a banner
• Keep an accelerator ledger — every late discovery that would have sped earlier work is recorded with when it was found, when it could have been, and what it would have saved
• Checkpoint means everything is already in a file — a checkpoint is the LAST thing written; a good commit message is not in the load order; the sweep before the block
• A PhaseEnd carries the narrative axis — what was believed, what failed and why it looked right, what it cost, what to do sooner — because transcripts die and the retrospective is rebuilt from the files
• Tool change ships with its consumers and docs — the commit that adds a tool also wires its callers and adds its ops row; integrating an agent's tool change IS a tool change; no end-of-session audits
• No sleep-polling background tasks — the harness re-invokes on completion; one sanity check after launch, then wait
• Long checks run in the foreground — a background long verification can be killed by the harness's memory guard; run it in the foreground with a timeout; check for orphaned workers
• Resume means resume the run — replay interrupted work so it finishes itself, same run, same agent, transcript intact; never re-plan or re-route on your own inference
• Answer before grinding in live co-op — only the final message of a turn is guaranteed to be seen; end the turn with the answer the person is waiting on before starting long work
• Consult the tool dictionary first — before designing or debugging a tool, look the NEED up in the tool dictionary; the proven implementation is the jumping-off point, its incident row the failure mode to avoid
• Translate the source idiom through the pass — an inherited idiom is a symptom + a named pass + a lever + a byte proof; same family: use and re-prove; other compiler: read the same pass in your compiler's source and find your own lever