feat: add re:factory persistent supervisor agent (#728)

[xukai92]

Summary

• New re:factory agent — a persistent Claude Code supervisor that manages CEO agent lifecycles, context/compaction, and playbook evolution
• Bare factory command (no subcommand) now launches the refactory agent instead of the welcome wizard
• 4 slash command skills: /factory-run (dispatch), /sessions (tracking), /compaction (context preservation), /playbook (ACE evolution)
• 13 new tests covering workspace setup, session management, role registration, and CLI integration

Files changed

• factory/agents/prompts/refactory.md — agent system prompt (149 lines)
• factory/refactory.py — workspace setup + session management module
• factory/agents/skills/ — 4 skill files (factory-run, sessions, compaction, playbook)
• factory/agents/runner.py — added "refactory" to AgentRole
• factory/agents/agents.yml — added refactory role config
• factory/cli.py — added cmd_refactory + wired bare factory to launch it
• tests/test_refactory.py — 13 test cases

Test plan

• uv run pytest tests/test_refactory.py -v — 13/13 pass
• uv run pytest tests/test_prompts.py -v — 71/71 pass (no regressions)
• ruff check factory/refactory.py factory/agents/runner.py — clean
• Full test suite (366 tests in affected files) — all pass

Closes #728

🤖 Generated with Claude Code

[github-actions]

Sentrux Quality Report

Absolute

Scanning ....
[scan] git ls-files: 312 total, 300 kept, 12 dropped (ext:12, meta:0, big:0)
[build_project_map] 300 files, 54 unique dirs, 50 cache misses, 2.3ms
[resolve] 444 resolved, 786 unresolved (of 1230 total specs)
[resolve_imports] project_map 2.4ms, suffix_idx 0.6ms, suffix_resolve 10.4ms, total 13.4ms
[build_graphs] 300 files | maps 0.9ms, imports 13.5ms, calls+inherit 3.1ms, total 17.4ms | 443 import, 4416 call, 0 inherit edges
sentrux check — 2 rules checked

Quality: 4705

✗ [Error] max_cc: 5 function(s) exceed max cyclomatic complexity of 30
    factory/cli.py:cmd_ceo (cc=79)
    factory/study.py:study_project_local (cc=43)
    factory/cli.py:_welcome_wizard (cc=39)
    factory/cli.py:cmd_run (cc=37)
    factory/workflow/validation.py:validate_workflow (cc=31)

✗ 1 violation(s) found

Diff (vs base branch)

Scanning ....
[scan] git ls-files: 312 total, 300 kept, 12 dropped (ext:12, meta:0, big:0)
[build_project_map] 300 files, 54 unique dirs, 50 cache misses, 2.4ms
[resolve] 444 resolved, 786 unresolved (of 1230 total specs)
[resolve_imports] project_map 2.5ms, suffix_idx 1.0ms, suffix_resolve 9.9ms, total 13.4ms
[build_graphs] 300 files | maps 0.8ms, imports 13.4ms, calls+inherit 3.1ms, total 17.4ms | 443 import, 4416 call, 0 inherit edges
sentrux gate — structural regression check

Quality:      4705 -> 4705
Coupling:     0.75 → 0.75
Cycles:       4 → 4
God files:    0 → 0

Distance from Main Sequence: 0.35

✓ No degradation detected

[codecov]

Codecov Report

❌ Patch coverage is 87.32394% with 9 lines in your changes missing coverage. Please review.
✅ Project coverage is 86.86%. Comparing base (3baba8c) to head (ac90ffe).
⚠️ Report is 37 commits behind head on main.

Files with missing lines	Patch %	Lines
factory/cli.py	67.85%	9 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #754      +/-   ##
==========================================
+ Coverage   86.75%   86.86%   +0.11%     
==========================================
  Files          79       81       +2     
  Lines       11881    12203     +322     
==========================================
+ Hits        10307    10600     +293     
- Misses       1574     1603      +29     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[xukai92]

@ceo-review

[github-actions]

✅ Factory Review: KEEP

Verdict: KEEP
Reason: Persistent refactory supervisor agent with full session management, context/compaction skills, and playbook evolution. All tests pass (2597/2597) including 19 new tests. 2 LOW severity issues are acceptable patterns. All acceptance criteria verified.

Score Comparison

Metric	Value
Before	0.4282
After	0.4282
Delta	+0.0000
Threshold	0.8000

Code Review Notes

• qa_verdict=ISSUES_FOUND
• qa_issues_count=2
• qa_issues_severity=LOW
• tests_pass=2597/2597
• lint_pass=yes
• acceptance_criteria=4/4
• issues=unreachable_wizard_code+temp_file_leak_with_execvp
• smoke_test=PASS
• edge_cases=9/9_PASS

---

Posted by Factory CEO