mirror of
https://github.com/garrytan/gstack.git
synced 2026-09-16 09:55:29 +02:00
feat(evals): with-skill vs without-skill arm benchmark — measures whether gstack's behavioral layer earns its tokens
Ponytail's honest-benchmark method pointed at gstack itself: 3 build-shaped
tasks (native-platform over-build trap, CRUD endpoint, bug fix with planted
decoys) x 2 arms, real claude -p sessions, scored on the git diff left
behind. A research instrument, not a release gate — no assertion compares
arm scores.
Arms use the PROVEN project-scope pattern: the with-arm installs a
build-discipline skill (extracted reuse-ladder + bounded-closer content, not
whole-file copies) into the fixture's .claude/skills/ with a CLAUDE.md
routing line and an explicit invocation; a live spike confirmed claude -p
discovers and invokes project-scope skills via the Skill tool (3 turns,
exact-output probe). Fixtures are git init + local bare origin; diff capture
is three lines of git, no worktree machinery.
Failure taxonomy: zero-diff arms are VALID scored cells (deterministic
0/none, no API call), harvest failures record harvest:null, judge_error
cells are excluded from aggregates but named in the report — nothing drops
silently. armJudge: fixed sonnet judge, 0-3 unrequested-structure rubric,
must name the construct or say none, bounded retry-on-malformed; callJudge
gains optional temperature/max_tokens (defaults unchanged). recordE2E now
populates tokens_used for every E2E. Eval schema v2: harvest gains
{insertions, deletions, net}, tolerant reads keep v1 runs comparable.
Registered periodic in E2E_TIERS + touchfiles (with the auq-repetition-cut
A/B); periodic detach timeout raised to the new shard-census floor. Free
selftest (8 tests, zero API) pins fixtures, extraction, arm asymmetry, diff
capture, judge plumbing, and the retry bound.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This commit is contained in:
co-authored by
Claude Fable 5
parent
781f46d025
commit
4c20eca33b
@@ -113,8 +113,10 @@ describe('selectTests', () => {
|
||||
expect(result.selected).toContain('plan-ceo-section-loading');
|
||||
// Token-reduction Phase 1: the preamble script A/B also keys on plan-ceo-review/**.
|
||||
expect(result.selected).toContain('preamble-script-ab');
|
||||
expect(result.selected.length).toBe(24);
|
||||
expect(result.skipped.length).toBe(Object.keys(E2E_TOUCHFILES).length - 24);
|
||||
// AUQ repetition-cut NOT-WORSE gate drives plan-ceo-review, so it keys on it too.
|
||||
expect(result.selected).toContain('auq-repetition-cut-ab');
|
||||
expect(result.selected.length).toBe(25);
|
||||
expect(result.skipped.length).toBe(Object.keys(E2E_TOUCHFILES).length - 25);
|
||||
});
|
||||
|
||||
test('global touchfile triggers ALL tests', () => {
|
||||
|
||||
Reference in New Issue
Block a user