learning-log

To document development discoveries during active coding, you can log bug root causes, workarounds, and surprising framework behaviors as structured entries with categories, timestamps, and affected-file references. This captures non-obvious findings before they are forgotten or misremembered.

Tracking architecture decisions and performance findings requires a searchable knowledge base that stores date, category, title, description, and affected files. This structure enables scanning and pattern detection to reuse organic development insights later instead of losing them.

Separating organic learning logs from corrective rules keeps non-obvious discoveries distinct from strict behavioral instructions. The learning log captures root causes and workarounds, while corrective rules and session state are reserved for separate memory stores to avoid mixing documentation types.

Record a workaround or external service quirk in a learning log when the causal explanation is non-obvious during a development session. Logging it immediately preserves the specific context and affected files so the discovery can be reused when the issue reappears.

A learning log entry does not automatically correct future code generation behavior. It explicitly avoids the role of corrective rules reserved for separate memory stores, functioning solely as a searchable record of organic discoveries, gotchas, and architecture decisions for later reference.