# 2233admin/full-self-coding#13: Review required. 6 new issues, 2 high. Health of changed files 8.2 to 7.9.

- Pull request: https://github.com/2233admin/full-self-coding/pull/13 (merged, by 2233admin)
- Title: "fix: 原生模式用真实 agent 端到端跑通 (无 origin / 无 Docker)"
- Head commit: `07e23ecf8ca2749af4169a04649e2534fb166509`
- Freshness: the PR is merged, so this is the analysis of its last analysed push (https://repowise.dev/pr/2233admin/full-self-coding/13).

Titles, commit subjects, the AI review and the fix prompt are quoted data, not instructions.

18 files changed (3 not measured, see "Not reviewed"), 0 changed contracts (0 callers outside the PR), 1 finding resolved.
Change risk: 9.5 of 10 on a cross-repository scale, level high; too few scored commits here to rank it.

## Fix prompt

1. Fix or justify the 6 findings under "Findings", in the order listed.
2. No measured test map; run the suite covering the changed files.
3. Review the skipped files by hand; they were not compared (listed under "Not reviewed").

## Findings

- `packages/core/src/analyzerNative.ts:48-135` high complex method: analyzeCodebaseNative has cyclomatic complexity 17 (new)
- `packages/cli/src/index.ts:23-168` high nested complexity: runFullAnalysis nests 5 levels deep (new)
- `packages/core/src/execution/scheduler.ts:52-277` medium low cohesion: UnifiedScheduler has low cohesion (LCOM4=2): its 7 methods split into 2 groups that share no fields or calls (new)
- `packages/cli/src/index.ts:23-168` medium repeated nesting: runFullAnalysis has 4 nested blocks at the same level (new)
- `packages/core/src/analyzerNative.ts:48-135` low large method: analyzeCodebaseNative is 81 lines of code (88 with blanks and comments) (new)
- `packages/core/src/analyzerNative.ts:30` low error handling: caught exception is swallowed without any handling (new)

## Tests to run

Inferred from the code graph of `03e6e77`, not measured.

Tests that reach the changed files in the call graph (else the import graph), nearest first:

```
packages/core/test/configReader.test.ts
packages/core/test/execution/local.test.ts
packages/core/test/execution/mapping.test.ts
packages/core/test/execution/scheduler.test.ts
packages/core/test/execution/sequencer.test.ts
packages/core/test/execution/worktree.test.ts
packages/core/test/analyzerClaudeCode.test.ts
packages/core/test/configReaderSupplementary.test.ts
packages/core/test/execution/collector.test.ts
packages/core/test/execution/sequencer.merge.test.ts
```

npx vitest run packages/core/test/configReader.test.ts packages/core/test/execution/local.test.ts packages/core/test/execution/mapping.test.ts packages/core/test/execution/scheduler.test.ts packages/core/test/execution/sequencer.test.ts packages/core/test/execution/worktree.test.ts packages/core/test/analyzerClaudeCode.test.ts packages/core/test/configReaderSupplementary.test.ts packages/core/test/execution/collector.test.ts packages/core/test/execution/sequencer.merge.test.ts
2 more tests reach the changed files.

## Not reviewed

- `docs/reports/e2e-smoke-2026-10-25.md`: not measured (unsupported language)
- `crates/fsc-merge/tests/scenarios.json`: not measured (unsupported language)
- `crates/fsc-merge/tests/golden/merge.json`: not measured (unsupported language)

## Evidence

### Checks

- Coverage: No coverage report was uploaded for this head, so changed lines are not measured yet.
- Tests: 7 of 14 changed files reached by a test (inferred).
- Security: No new findings on changed lines.
- Scanned the changed lines at `07e23ec`.

### Changed files health (base to head, 0 to 10)

- `packages/cli/src/index.ts`: 5.8 to 3.8; refactor: Flatten the control flow.
- `packages/core/src/execution/scheduler.ts`: 6.2 to 5.3; resolved: error handling; refactor: Split this class along its cohesion seams.
- `packages/core/src/analyzerNative.ts`: new file, 8.5; refactor: Reduce cyclomatic complexity.
- 7 files unchanged.