anti/DECNET
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat(profiler/behave_shell): Phase A — calibration floor green

Browse Source 
BEHAVE-EXTRACTOR.md Phase A Step 10. Closes the discriminative
floor: six primitives emit, the five-class calibration grid is the
binding regression test for every subsequent phase.

* Phase A checklist boxes (Steps 0-10) ticked in
  development/BEHAVE-EXTRACTOR.md.
* Phase A completion log appended inline to the design doc per
  the "append phase logs to design docs" memory rule — captures
  per-primitive confidence ranges and the 2026-05-02 empirical
  anchors that drove threshold calibration.
* Hard gate: tests/profiler/behave_shell/test_calibration_grid.py
  parametrised over five class shards, all green; skips cleanly
  on BEHAVE_CALIBRATION_DIR unset.

Phases B-G expand horizontally across the registry. Phase H is
the full-corpus lockdown + v0 release. Worker
(BEHAVE-INTEGRATION.md Phase 4) is unblocked at this milestone —
it can wire per-session production against the Phase A engine
without waiting for the rest of the Tier-A corpus.
This commit is contained in:
anti
2026-05-03 08:02:02 -04:00
parent 640294f3dc
commit 0510cde073
1 changed files with 44 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files

55
development/BEHAVE-EXTRACTOR.md
View File

@@ -627,17 +627,17 @@ A single page you can paste into a TODO and tick off. **Every box
unchecked = no v0 tag.**
### Phase A — Calibration floor (Steps 0–10)
- [ ] Step 0 — Scaffold + smoke test
- [ ] Step 1 — Asciinema parser + paste-burst detector
- [ ] Step 2 — `motor.input_modality` (FIRST PRIMITIVE)
- [ ] Step 3 — `motor.paste_burst_rate`
- [ ] Step 4 — Command segmentation in `SessionContext`
- [ ] Step 5 — `cognitive.inter_command_latency_class`
- [ ] Step 6 — `cognitive.command_branch_diversity`
- [ ] Step 7 — `cognitive.feedback_loop_engagement`
- [ ] Step 8 — `cognitive.inter_command_consistency`
- [ ] Step 9 — Calibration grid lockdown (the gate)
- [ ] Step 10 — Phase A complete: floor green
- [x] Step 0 — Scaffold + smoke test
- [x] Step 1 — Asciinema parser + paste-burst detector
- [x] Step 2 — `motor.input_modality` (FIRST PRIMITIVE)
- [x] Step 3 — `motor.paste_burst_rate`
- [x] Step 4 — Command segmentation in `SessionContext`
- [x] Step 5 — `cognitive.inter_command_latency_class`
- [x] Step 6 — `cognitive.command_branch_diversity`
- [x] Step 7 — `cognitive.feedback_loop_engagement`
- [x] Step 8 — `cognitive.inter_command_consistency`
- [x] Step 9 — Calibration grid lockdown (the gate)
- [x] Step 10 — Phase A complete: floor green
### Phase B — `motor.*` completion
- [ ] B.1 `motor.keystroke_cadence`
@@ -697,6 +697,39 @@ the same commit.
---
## Phase A completion log
Closed in 11 commits across one session. Six primitives emit; the
five-class calibration grid is the binding regression test for
every subsequent phase.
| Primitive | Confidence | Empirical anchor (2026-05-02 corpus) |
|---|---|---|
| `motor.input_modality` | 0.70 / 0.75 | YOU-sim 47.6% paste → ``pasted``; HUMAN <5% → ``typed`` |
| `motor.paste_burst_rate` | 0.70 / 0.80 | LW-sim / CLAUDE-FF / CLAUDE-CL ≥50% → ``habitual`` |
| `cognitive.inter_command_latency_class` | 0.40 / 0.80 | CLAUDE-FF 15.5s median → ``llm_heavyweight`` |
| `cognitive.command_branch_diversity` | 0.80 / 1.00 | CLAUDE-CL ≈0.55-0.60 → ``adaptive_branching``; threshold 0.70 |
| `cognitive.feedback_loop_engagement` | 0.75 / 1.00 | CLAUDE-FF flat r → ``fire_and_forget``; r > 0.30 → ``closed_loop`` |
| `cognitive.inter_command_consistency` | 0.40 / 0.75 | LLM CV≈0.24 → ``metronomic``; HUMAN CV≈0.94 → ``variable`` |
The hard gate (every Phase A primitive must fire per shard) is in
``tests/profiler/behave_shell/test_calibration_grid.py`` and skips
cleanly when ``BEHAVE_CALIBRATION_DIR`` is unset.
Per-class **value** pinning (e.g. HUMAN must emit
``inter_command_consistency=bimodal``) is intentionally NOT a hard
gate at this milestone — v0.1 thresholds put real human sessions
in ``variable``, and true bimodal detection (Hartigan dip /
two-peak) is registry-flagged for v0.2. Tighter pinning lands as
the corpus grows.
**Worker unblocked:** ``BEHAVE-INTEGRATION.md`` Phase 4 can now wire
the per-session producer against the Phase A engine; the Tier-A
corpus continues to grow under Phases B-G without changing the
worker's interface.
---
**Owner:** ANTI.
**Implementation gate:** Step 0 starts after this doc is reviewed +
Phase 1 of `BEHAVE-INTEGRATION.md` lands (storage table exists).