Earlier amendment investigations and run evidence

These files retain the October 2–3 ID investigation, completed repair checks and Dynamic Cheatsheet audits. They describe their original method boundary; they are evidence, not the current plan or a complete description of the live method. Start new design work at the workshop README.

ID investigation and adoption

  • Problem — starting evidence and record-identity questions.
  • Fix options — considered repairs and their dispositions.
  • Split-records proposal — the adopted Part of: repair at that time, not a new pending proposal.
  • Pre-adoption check — normalized archive copies, part relations, negative checks and bounded workflow fixtures. Links to the two copied sets under rewrite-check/ are in this check.
  • Implementation check — adopted repairs, consumers, verification and original stopping points.
  • Skill inconsistencies — audit of the method before those repairs; its suggested order is not a current work queue.

Dynamic Cheatsheet runs

The operator accepted the second result unchanged on 2026-10-03. Its remaining findings are comparison evidence. Their repair was deferred until recurrence; these files do not commission another repair or rerun.

Earlier framing and completed plans

The full earlier commission, implementation chronology and completed minimal pre-run plan remain available through git:

git show 1b5a5a8b7:kb/work/analyse-agentic-system-amendments/README.md

Use that read path to reconstruct historical authority or change sequence. Use the live code and contracts for current behavior.

Validation boundary

The eight rewrite-check/ files are byte-identical to their pre-cleanup versions. Their historical type: pointers name retired global specs: types/agentic-system-epistemic-report.md, types/agent-memory-analysis-report.md, types/agentic-system-analysis-overview.md and types/agentic-system-runtime-report.md. Directory validation reports eight missing-type-spec failures. Do not silently migrate these copied fixtures to current schemas and treat the result as the original acceptance evidence. The surrounding history and current workshop documents validate separately.

Retention boundary

This is local workshop history, not a permanent archive. At closure, extract only evidence with an independent durable consumer, preserve any needed follow-up, and delete the rest with the workshop. The Sol follow-up plan remains outside history until its unresolved dispositions are checked.

Complete file listing (generated at build time)