feat: enhance optimize-sorting T2 researcher and archive prompts with stage timing

[colehurwitz]

Changes

• T2 Researcher Prompt (_RESEARCHER_T2_PROMPT): Updated to instruct the researcher to read .stage_timing from baseline.json as ground-truth measured times and rank stages by measured time (highest first), replacing the old "estimate from code" approach. Output schema changed from estimated_pct to baseline_mean/baseline_std/pct_of_total.

• Archive Prompt (_archive_prompt(), all 3 tiers): Added stage_timing_deltas computation that iterates over the union of baseline and result stage names:

  • Stages in both → normal delta (baseline_mean - result_time)
  • Stages only in result → marked "new"
  • Stages only in baseline → marked "removed"
  • Added stage_timing_deltas to JSONL field list
  • Added per-stage timing table to PR comment for keep verdicts

• Tests: Added 10 new test methods to TestPromptContent validating both changes

Graph topology unchanged

• 31 nodes, 39 edges, 2 reloop edges — verified programmatically
• All existing structural tests pass (98/98 non-registration tests)
• 3 registration tests expected to fail (workflow not yet registered in definitions.py)

[github-actions]

Sentrux Quality Report

Absolute

Scanning ....
[scan] git ls-files: 629 total, 615 kept, 14 dropped (ext:14, meta:0, big:0)
[build_project_map] 615 files, 109 unique dirs, 101 cache misses, 5.0ms
[resolve] 1165 resolved, 1572 unresolved (of 2737 total specs)
[resolve_imports] project_map 5.1ms, suffix_idx 1.0ms, suffix_resolve 14.0ms, total 20.2ms
[build_graphs] 615 files | maps 1.9ms, imports 20.3ms, calls+inherit 4.5ms, total 26.7ms | 1164 import, 8401 call, 1 inherit edges
sentrux check — 3 rules checked

Quality: 4450

✗ [Error] max_cc: 6 function(s) exceed max cyclomatic complexity of 30
    factory/cli/_ceo_helpers.py:_validate_ceo_flags (cc=43)
    factory/cli/_ceo_helpers.py:_execute_ceo (cc=43)
    examples/chess_evolve.py:main (cc=33)
    factory/cli/_task_builder.py:_build_ceo_task (cc=32)
    factory/cli/run.py:cmd_run (cc=32)
    factory/outer_loop/mutations.py:validate_and_repair (cc=31)

✗ 1 violation(s) found

Diff (vs base branch)

Scanning ....
[scan] git ls-files: 629 total, 615 kept, 14 dropped (ext:14, meta:0, big:0)
[build_project_map] 615 files, 109 unique dirs, 101 cache misses, 5.1ms
[resolve] 1165 resolved, 1572 unresolved (of 2737 total specs)
[resolve_imports] project_map 5.2ms, suffix_idx 1.0ms, suffix_resolve 14.1ms, total 20.3ms
[build_graphs] 615 files | maps 1.9ms, imports 20.5ms, calls+inherit 5.8ms, total 28.1ms | 1164 import, 8401 call, 1 inherit edges
sentrux gate — structural regression check

Quality:      4447 -> 4450
Coupling:     0.80 → 0.80
Cycles:       4 → 4
God files:    3 → 3

Distance from Main Sequence: 0.35

✓ No degradation detected

[colehurwitz]

Superseded by unified PR — see #1443