Thanks for the work here. There is useful progress in this branch, but this PR should be split before merge.

Right now it is too broad to review safely: 28 changed files, ~4k additions, and several distinct concerns bundled together. In particular, it mixes:

1. Core workflow state changes (run_manifest, rollback, stale/dirty stage tracking, CLI flags)
2. Operator/session recovery and stage handoff/compression logic
3. Stage 07/08 publication-package changes
4. A large new src/platform/* stack plus knowledge_base.py / inspection.py
5. README + test expansion

These are not one review unit. Some are core workflow changes, some are reliability improvements, some are writing-package features, and some are a substantial architectural expansion. Reviewing them together makes it hard to reason about regressions, approve only the good parts, or maintain a clear project direction.

Suggested split:

• PR A: workflow-state layer only

  • main.py, src/manager.py, src/manifest.py, src/run_state.py, src/utils.py
  • focus on run_manifest, rollback, stale downstream invalidation, and state transitions

• PR B: operator recovery / continuation only

  • src/operator.py + the minimal related manager changes + targeted tests
  • focus on session recovery, failed resume fallback, attempt metadata, handoff/compression if still needed

• PR C: Stage 07/08 packaging only

  • publication package, review/dissemination artifacts, README updates relevant to that scope, and tests for that slice

• PR D: platform modules only, if they are still desired

  • src/platform/*, src/knowledge_base.py, src/inspection.py
  • this is a major architectural addition by itself and should be reviewed independently from the CLI workflow changes

Please keep each split PR narrowly scoped, with its own motivation, tests, and validation. In the current form, this is too much surface area for one merge.

@yyifan-onyan

Code Review

I agree with @black-yt's suggestion to split this PR — 28 files and ~4k additions across multiple unrelated concerns is too much surface area for a single review.

Beyond the split, there's a larger issue: most of the core functionality in this PR has already landed on main through other PRs.

Already on main

These portions would conflict heavily on rebase and would essentially be duplicated work.

What's actually new

• --show-status and --kb-search CLI commands — useful, worth a focused PR
• src/knowledge_base.py and src/inspection.py — potentially useful but need their own review
• src/platform/* (14 files) — see below

Concerns about src/platform/*

Many of the 14 platform modules are architectural stubs rather than functional code:

• sandbox.py (39 lines): SandboxRunner just calls subprocess.run directly — no actual sandboxing
• security.py (55 lines): RBAC role definitions, but AutoR is a single-user CLI tool
• messaging.py (31 lines), protocols.py (53 lines): interface definitions with no implementation
• semantic.py (52 lines): token-overlap ranking presented as "semantic search" — not embedding-based

The modules with real substance (router.py 269 lines, literature.py 271 lines, debate.py 157 lines) are imported by manager.py but not actually used in the core _run_stage loop — the router is instantiated but the stage execution still goes through the existing ClaudeOperator path.

Merge conflicts

8 files currently conflict with main: README.md, main.py, src/manager.py, src/manifest.py, src/operator.py, src/platform/foundry.py, src/utils.py, tests/test_operator_recovery.py.

Suggested path forward

1. Drop the already-merged portions (manifest, rollback, operator recovery, handoff) — they're on main already
2. PR A: --show-status + --kb-search CLI commands with knowledge_base.py — small, reviewable, useful
3. PR B: Platform modules that have real functionality (router, literature, debate) — but only if they're wired into the actual stage execution, not just imported
4. Hold off on stub modules (sandbox, security, messaging, protocols, semantic) until there's a concrete use case driving them

Thank you for your contribution. This PR has been superseded by your newer PR, so we are closing it.