@lythos/skill-arena 0.16.0 → 0.16.1

package/README.md

@@ -16,20 +16,20 @@
 ```bash
 bun add -d @lythos/skill-arena
 # or use directly
-bunx @lythos/skill-arena@0.16.0 <command>
+bunx @lythos/skill-arena@0.16.1 <command>
 ```
 
 ## Quick Start
 
 ```bash
 # single — test one deck (most common)
-bunx @lythos/skill-arena@0.16.0 single \
+bunx @lythos/skill-arena@0.16.1 single \
   --deck ./examples/decks/scout.toml \
   --brief "Generate auth flow diagram" \
   --out ./output
 
 # single with explicit player
-bunx @lythos/skill-arena@0.16.0 single \
+bunx @lythos/skill-arena@0.16.1 single \
   --deck ./examples/decks/scout.toml \
   --brief "Generate auth flow diagram" \
   --player kimi \
@@ -37,10 +37,10 @@ bunx @lythos/skill-arena@0.16.0 single \
 
 # cross-deck vs — compare two decks (agent-orchestrated)
 # Create arena.toml declaring sides with different decks, then:
-bunx @lythos/skill-arena@0.16.0 vs --config ./arena.toml
+bunx @lythos/skill-arena@0.16.1 vs --config ./arena.toml
 
 # cross-player vs — compare kimi vs codex (CLI only)
-bunx @lythos/skill-arena@0.16.0 vs --config ./arena.toml --player kimi
+bunx @lythos/skill-arena@0.16.1 vs --config ./arena.toml --player kimi
 ```
 
 **What happens**: Agent creates isolated `/tmp` workdir per side, `deck link` skills, spawns parallel subagents, collects artifacts, judge scores outputs. Parent deck restored after.
@@ -50,14 +50,14 @@ bunx @lythos/skill-arena@0.16.0 vs --config ./arena.toml --player kimi
 ### `single` — one deck, one task
 
 ```bash
-bunx @lythos/skill-arena@0.16.0 single \
+bunx @lythos/skill-arena@0.16.1 single \
   --deck ./deck.toml \
   --brief "Produce a .docx report with radar chart" \
   --timeout 600000 \
   --out ./output
 
 # with explicit player
-bunx @lythos/skill-arena@0.16.0 single \
+bunx @lythos/skill-arena@0.16.1 single \
   --deck ./deck.toml \
   --brief "Produce a .docx report with radar chart" \
   --player kimi \
@@ -67,14 +67,14 @@ bunx @lythos/skill-arena@0.16.0 single \
 ### `vs` — multi-deck comparison
 
 ```bash
-bunx @lythos/skill-arena@0.16.0 vs --config ./arena.toml
-bunx @lythos/skill-arena@0.16.0 vs --config ./arena.toml --dry-run
+bunx @lythos/skill-arena@0.16.1 vs --config ./arena.toml
+bunx @lythos/skill-arena@0.16.1 vs --config ./arena.toml --dry-run
 ```
 
 ### `prepare-workdir` — isolate + link skills (agent-orchestrated)
 
 ```bash
-bunx @lythos/skill-arena@0.16.0 prepare-workdir \
+bunx @lythos/skill-arena@0.16.1 prepare-workdir \
   --deck ./skill-deck.toml \
   --out /tmp/arena-side-a \
   --brief "task description"
@@ -85,7 +85,7 @@ Creates `/tmp`-isolated workdir with deck copied, AGENTS.md written, and `deck l
 ### `archive` — collect agent outputs (agent-orchestrated)
 
 ```bash
-bunx @lythos/skill-arena@0.16.0 archive \
+bunx @lythos/skill-arena@0.16.1 archive \
   --from /tmp/arena-side-a \
   --to ./playground/output \
   --sides side-a
@@ -96,7 +96,7 @@ Copies agent artifacts from workdir(s) to output, skipping internal files (`.cla
 ### `viz` — render results (WIP — HTML report generation pending)
 
 ```bash
-bunx @lythos/skill-arena@0.16.0 viz runs/arena-<id>/
+bunx @lythos/skill-arena@0.16.1 viz runs/arena-<id>/
 ```
 
 ## Parameters

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@lythos/skill-arena",
-  "version": "0.16.0",
+  "version": "0.16.1",
   "description": "Skill Arena — benchmark skill effectiveness with controlled-variable comparison",
   "keywords": [
     "ai-agent",
@@ -42,15 +42,15 @@
     "bun": ">=1.0.0"
   },
   "dependencies": {
-    "@lythos/cold-pool": "^0.16.0",
-    "@lythos/infra": "^0.16.0",
-    "@lythos/test-utils": "^0.16.0",
+    "@lythos/cold-pool": "^0.16.1",
+    "@lythos/infra": "^0.16.1",
+    "@lythos/test-utils": "^0.16.1",
     "zod": "^3.24.0",
     "zod-to-json-schema": "^3.25.2"
   },
   "optionalDependencies": {
-    "@lythos/agent-adapter-claude-sdk": "^0.16.0",
-    "@lythos/agent-adapter-deepseek-serve": "^0.16.0",
-    "@lythos/agent-adapter-codex": "^0.16.0"
+    "@lythos/agent-adapter-claude-sdk": "^0.16.1",
+    "@lythos/agent-adapter-deepseek-serve": "^0.16.1",
+    "@lythos/agent-adapter-codex": "^0.16.1"
   }
 }

package/src/arena-toml.test.ts

@@ -1,4 +1,4 @@
-import { describe, test, expect } from 'bun:test'
+import { describe, it, expect } from 'bun:test'
 import { parseArenaToml, buildExecutionPlan, ArenaToml } from './arena-toml'
 import { formatPlanOutput } from './runner'
 
@@ -63,7 +63,7 @@ working_dir = "/workspace"
 `
 
 describe('parseArenaToml', () => {
-  test('parses minimal two-side arena with criteria', () => {
+  it('parses minimal two-side arena with criteria', () => {
     const result = parseArenaToml(minimalToml)
     expect(result.arena.task).toBe('Test task')
     expect(result.arena.criteria).toEqual(['a', 'b'])
@@ -75,14 +75,14 @@ describe('parseArenaToml', () => {
     expect(result.side[0].control).toBe(false)
   })
 
-  test('parses arena with judge field (preferred over criteria)', () => {
+  it('parses arena with judge field (preferred over criteria)', () => {
     const result = parseArenaToml(judgeToml)
     expect(result.arena.judge).toContain('Evaluate completeness')
     expect(result.arena.criteria).toBeUndefined()
     expect(result.side).toHaveLength(2)
   })
 
-  test('parses full arena with runs_per_side and control', () => {
+  it('parses full arena with runs_per_side and control', () => {
     const result = parseArenaToml(fullToml)
     expect(result.arena.runs_per_side).toBe(3)
     expect(result.side).toHaveLength(3)
@@ -90,7 +90,7 @@ describe('parseArenaToml', () => {
     expect(result.side[2].control).toBe(true)
   })
 
-  test('parses side env block', () => {
+  it('parses side env block', () => {
     const result = parseArenaToml(fullToml)
     const env = result.side[2].env
     expect(env.container).toBe('node:20-alpine')
@@ -99,47 +99,47 @@ describe('parseArenaToml', () => {
     expect(env.env_vars).toEqual({})
   })
 
-  test('rejects fewer than 2 sides', () => {
+  it('rejects fewer than 2 sides', () => {
     const bad = `[arena]\ntask = "x"\ncriteria = ["a"]\n\n[[side]]\nname = "only"\nplayer = "c"\ndeck = "x.toml"`
     expect(() => parseArenaToml(bad)).toThrow()
   })
 
-  test('rejects neither judge nor criteria provided', () => {
+  it('rejects neither judge nor criteria provided', () => {
     const bad = `[arena]\ntask = "x"\n\n[[side]]\nname = "a"\nplayer = "c"\ndeck = "a.toml"\n\n[[side]]\nname = "b"\nplayer = "c"\ndeck = "b.toml"`
     expect(() => parseArenaToml(bad)).toThrow()
   })
 
-  test('accepts judge without criteria (either is sufficient)', () => {
+  it('accepts judge without criteria (either is sufficient)', () => {
     const toml = `[arena]\ntask = "x"\njudge = "Evaluate this."\n\n[[side]]\nname = "a"\nplayer = "c"\ndeck = "a.toml"\n\n[[side]]\nname = "b"\nplayer = "c"\ndeck = "b.toml"`
     expect(() => parseArenaToml(toml)).not.toThrow()
   })
 
-  test('rejects empty criteria and no judge', () => {
+  it('rejects empty criteria and no judge', () => {
     const bad = `[arena]\ntask = "x"\ncriteria = []\n\n[[side]]\nname = "a"\nplayer = "c"\ndeck = "a.toml"\n\n[[side]]\nname = "b"\nplayer = "c"\ndeck = "b.toml"`
     expect(() => parseArenaToml(bad)).toThrow()
   })
 
-  test('rejects non-object input', () => {
+  it('rejects non-object input', () => {
     expect(() => ArenaToml.parse('not valid')).toThrow()
   })
 
-  test('rejects missing arena section', () => {
+  it('rejects missing arena section', () => {
     expect(() => parseArenaToml('[[side]]\nname = "a"')).toThrow()
   })
 
-  test('rejects runs_per_side = 0', () => {
+  it('rejects runs_per_side = 0', () => {
     const bad = `[arena]\ntask = "x"\ncriteria = ["a"]\nruns_per_side = 0\n\n[[side]]\nname = "a"\nplayer = "c"\ndeck = "a.toml"\n\n[[side]]\nname = "b"\nplayer = "c"\ndeck = "b.toml"`
     expect(() => parseArenaToml(bad)).toThrow()
   })
 
-  test('parses integer and boolean values correctly', () => {
+  it('parses integer and boolean values correctly', () => {
     const toml = `[arena]\ntask = "x"\ncriteria = ["a"]\nruns_per_side = 2\nmax_participants = 5\n\n[[side]]\nname = "a"\nplayer = "c"\ndeck = "a.toml"\n\n[[side]]\nname = "b"\nplayer = "c"\ndeck = "b.toml"`
     const result = parseArenaToml(toml)
     expect(result.arena.runs_per_side).toBe(2)
     expect(result.arena.max_participants).toBe(5)
   })
 
-  test('comments are stripped', () => {
+  it('comments are stripped', () => {
     const toml = `[arena]\n# this is a comment\ntask = "x"\ncriteria = ["a"]\n\n[[side]]\nname = "a"\nplayer = "c"\ndeck = "a.toml"\n\n[[side]]\nname = "b"\nplayer = "c"\ndeck = "b.toml"`
     const result = parseArenaToml(toml)
     expect(result.arena.task).toBe('x')
@@ -147,7 +147,7 @@ describe('parseArenaToml', () => {
 })
 
 describe('buildExecutionPlan', () => {
-  test('generates plan: 2 sides × 1 run = 2 cells', () => {
+  it('generates plan: 2 sides × 1 run = 2 cells', () => {
     const toml = parseArenaToml(minimalToml)
     const plan = buildExecutionPlan(toml)
     expect(plan.task).toBe('Test task')
@@ -156,20 +156,20 @@ describe('buildExecutionPlan', () => {
     expect(plan.total_runs).toBe(2)
   })
 
-  test('generates plan with judge field populated', () => {
+  it('generates plan with judge field populated', () => {
     const toml = parseArenaToml(judgeToml)
     const plan = buildExecutionPlan(toml)
     expect(plan.judge).toContain('Evaluate completeness')
   })
 
-  test('generates plan: 3 sides × 3 runs = 9 cells', () => {
+  it('generates plan: 3 sides × 3 runs = 9 cells', () => {
     const toml = parseArenaToml(fullToml)
     const plan = buildExecutionPlan(toml)
     expect(plan.cells).toHaveLength(9)
     expect(plan.total_runs).toBe(9)
   })
 
-  test('control flag preserved in plan cells', () => {
+  it('control flag preserved in plan cells', () => {
     const toml = parseArenaToml(fullToml)
     const plan = buildExecutionPlan(toml)
     const baselineCells = plan.cells.filter(c => c.side === 'baseline')
@@ -177,7 +177,7 @@ describe('buildExecutionPlan', () => {
     expect(baselineCells.every(c => c.control)).toBe(true)
   })
 
-  test('dry-run: plan is pure data, no side effects', () => {
+  it('dry-run: plan is pure data, no side effects', () => {
     const toml = parseArenaToml(fullToml)
     const plan = buildExecutionPlan(toml)
     expect(plan.total_runs).toBeGreaterThan(0)

package/src/comparative-judge.test.ts

@@ -1,10 +1,10 @@
-import { describe, test, expect } from 'bun:test'
+import { describe, it, expect } from 'bun:test'
 import { computePareto, buildComparativePrompt, toScoreMatrix, normalizeComparativeOutput } from './comparative-judge'
 import { ArenaManifest, CriterionDef, ComparativeReport } from '@lythos/test-utils/schema'
 import type { ArenaManifest as ArenaManifestType } from '@lythos/test-utils/schema'
 
 describe('computePareto', () => {
-  test('single participant is always non-dominated', () => {
+  it('single participant is always non-dominated', () => {
     const result = computePareto([
       { participant_id: 'run-01', scores: { a: 5, b: 3 } },
     ])
@@ -13,7 +13,7 @@ describe('computePareto', () => {
     expect(result[0].dominated_by).toEqual([])
   })
 
-  test('clear dominance: run-01 dominates run-02 on all criteria', () => {
+  it('clear dominance: run-01 dominates run-02 on all criteria', () => {
     const result = computePareto([
       { participant_id: 'run-01', scores: { coverage: 5, relevance: 5 } },
       { participant_id: 'run-02', scores: { coverage: 3, relevance: 2 } },
@@ -23,7 +23,7 @@ describe('computePareto', () => {
     expect(result[1].dominated_by).toEqual(['run-01'])
   })
 
-  test('equal scores: no one dominates', () => {
+  it('equal scores: no one dominates', () => {
     const result = computePareto([
       { participant_id: 'run-01', scores: { a: 4, b: 4 } },
       { participant_id: 'run-02', scores: { a: 4, b: 4 } },
@@ -32,7 +32,7 @@ describe('computePareto', () => {
     expect(result[1].dominated).toBe(false)
   })
 
-  test('cross dominance: each wins on different criteria', () => {
+  it('cross dominance: each wins on different criteria', () => {
     const result = computePareto([
       { participant_id: 'run-01', scores: { speed: 5, accuracy: 2 } },
       { participant_id: 'run-02', scores: { speed: 2, accuracy: 5 } },
@@ -42,7 +42,7 @@ describe('computePareto', () => {
     expect(result[1].dominated).toBe(false)
   })
 
-  test('multi-participant: transitive dominance chain', () => {
+  it('multi-participant: transitive dominance chain', () => {
     const result = computePareto([
       { participant_id: 'best', scores: { a: 5, b: 5, c: 5 } },
       { participant_id: 'mid', scores: { a: 4, b: 4, c: 4 } },
@@ -56,7 +56,7 @@ describe('computePareto', () => {
     expect(result[2].dominated_by.sort()).toEqual(['best', 'mid'].sort())
   })
 
-  test('Pareto frontier from playground BDD-research: run-01 dominates run-02', () => {
+  it('Pareto frontier from playground BDD-research: run-01 dominates run-02', () => {
     // From playground/arena-bdd-research/report.md:
     // Run-01: coverage=5, relevance=5, actionability=5, depth=5
     // Run-02: coverage=3, relevance=2, actionability=2, depth=1
@@ -69,7 +69,7 @@ describe('computePareto', () => {
     expect(result[1].dominated_by).toEqual(['run-01'])
   })
 
-  test('empty scores object', () => {
+  it('empty scores object', () => {
     const result = computePareto([
       { participant_id: 'a', scores: {} },
       { participant_id: 'b', scores: {} },
@@ -79,7 +79,7 @@ describe('computePareto', () => {
     expect(result[1].dominated).toBe(false)
   })
 
-  test('partial criteria overlap', () => {
+  it('partial criteria overlap', () => {
     const result = computePareto([
       { participant_id: 'run-01', scores: { a: 5, b: 3 } },
       { participant_id: 'run-02', scores: { a: 3, c: 5 } },
@@ -103,7 +103,7 @@ const manifestFixture: ArenaManifest = {
 }
 
 describe('buildComparativePrompt', () => {
-  test('includes task description', () => {
+  it('includes task description', () => {
     const prompt = buildComparativePrompt({
       manifest: manifestFixture,
       verdicts: [],
@@ -111,7 +111,7 @@ describe('buildComparativePrompt', () => {
     expect(prompt).toContain('Write a function that adds two numbers')
   })
 
-  test('includes all participants', () => {
+  it('includes all participants', () => {
     const prompt = buildComparativePrompt({
       manifest: manifestFixture,
       verdicts: [],
@@ -122,7 +122,7 @@ describe('buildComparativePrompt', () => {
     expect(prompt).toContain('Full test discipline')
   })
 
-  test('includes criteria list', () => {
+  it('includes criteria list', () => {
     const prompt = buildComparativePrompt({
       manifest: manifestFixture,
       verdicts: [],
@@ -131,7 +131,7 @@ describe('buildComparativePrompt', () => {
     expect(prompt).toContain('efficiency')
   })
 
-  test('includes Zod schema in output spec', () => {
+  it('includes Zod schema in output spec', () => {
     const prompt = buildComparativePrompt({
       manifest: manifestFixture,
       verdicts: [],
@@ -145,7 +145,7 @@ describe('buildComparativePrompt', () => {
 // ── toScoreMatrix (pure Zod validation wrapper) ──────────────────────
 
 describe('toScoreMatrix', () => {
-  test('passes through valid score cells', () => {
+  it('passes through valid score cells', () => {
     const result = toScoreMatrix(manifestFixture, [
       { participant_id: 'bare', criterion: 'correctness', weight: 0.5, score: 4, rationale: 'works' },
       { participant_id: 'bare', criterion: 'efficiency', weight: 0.5, score: 3, rationale: 'ok' },
@@ -166,7 +166,7 @@ const sampleScoreMatrix = [
 ]
 
 describe('normalizeComparativeOutput', () => {
-  test('passes through already-correct format', () => {
+  it('passes through already-correct format', () => {
     const input = {
       score_matrix: sampleScoreMatrix,
       key_findings: ['TDD produced cleaner code'],
@@ -177,7 +177,7 @@ describe('normalizeComparativeOutput', () => {
     expect((result.score_matrix as any[])[0].participant_id).toBe('bare')
   })
 
-  test('maps participantId to participant_id', () => {
+  it('maps participantId to participant_id', () => {
     const input: Record<string, unknown> = {
       score_matrix: [
         { participantId: 'bare', criterion: 'accuracy', weight: 0.5, score: 4, rationale: 'good' },
@@ -189,7 +189,7 @@ describe('normalizeComparativeOutput', () => {
     expect((result.score_matrix as any[])[0].participant_id).toBe('bare')
   })
 
-  test('maps side to participant_id', () => {
+  it('maps side to participant_id', () => {
     const input: Record<string, unknown> = {
       score_matrix: [
         { side: 'tdd', criterion: 'quality', weight: 0.5, score: 5, rationale: 'excellent' },
@@ -201,7 +201,7 @@ describe('normalizeComparativeOutput', () => {
     expect((result.score_matrix as any[])[0].participant_id).toBe('tdd')
   })
 
-  test('coerces string score to number', () => {
+  it('coerces string score to number', () => {
     const input: Record<string, unknown> = {
       score_matrix: [
         { participant_id: 'bare', criterion: 'a', weight: 0.5, score: '4', rationale: 'ok' },
@@ -213,7 +213,7 @@ describe('normalizeComparativeOutput', () => {
     expect((result.score_matrix as any[])[0].score).toBe(4)
   })
 
-  test('normalizes weight >1 as percentage', () => {
+  it('normalizes weight >1 as percentage', () => {
     const input: Record<string, unknown> = {
       score_matrix: [
         { participant_id: 'bare', criterion: 'a', weight: 50, score: 4, rationale: 'ok' },
@@ -225,7 +225,7 @@ describe('normalizeComparativeOutput', () => {
     expect((result.score_matrix as any[])[0].weight).toBe(0.5)
   })
 
-  test('defaults weight to 0.25 when undefined', () => {
+  it('defaults weight to 0.25 when undefined', () => {
     const input: Record<string, unknown> = {
       score_matrix: [
         { participant_id: 'bare', criterion: 'a', score: 4, rationale: 'ok' },
@@ -237,7 +237,7 @@ describe('normalizeComparativeOutput', () => {
     expect((result.score_matrix as any[])[0].weight).toBe(0.25)
   })
 
-  test('maps reason to rationale', () => {
+  it('maps reason to rationale', () => {
     const input: Record<string, unknown> = {
       score_matrix: [
         { participant_id: 'bare', criterion: 'a', weight: 0.5, score: 4, reason: 'looks fine' },
@@ -249,7 +249,7 @@ describe('normalizeComparativeOutput', () => {
     expect((result.score_matrix as any[])[0].rationale).toBe('looks fine')
   })
 
-  test('maps explanation to rationale', () => {
+  it('maps explanation to rationale', () => {
     const input: Record<string, unknown> = {
       score_matrix: [
         { participant_id: 'bare', criterion: 'a', weight: 0.5, score: 4, explanation: 'works' },
@@ -261,7 +261,7 @@ describe('normalizeComparativeOutput', () => {
     expect((result.score_matrix as any[])[0].rationale).toBe('works')
   })
 
-  test('normalizes recommendations with role fallback', () => {
+  it('normalizes recommendations with role fallback', () => {
     const input: Record<string, unknown> = {
       score_matrix: [],
       key_findings: [],
@@ -275,7 +275,7 @@ describe('normalizeComparativeOutput', () => {
     expect(recs[0].recommendation).toBe('Use TDD')
   })
 
-  test('normalizes recommendations with advice fallback', () => {
+  it('normalizes recommendations with advice fallback', () => {
     const input: Record<string, unknown> = {
       score_matrix: [],
       key_findings: [],
@@ -287,7 +287,7 @@ describe('normalizeComparativeOutput', () => {
     expect((result.recommendations as any[])[0].recommendation).toBe('Consider refactoring')
   })
 
-  test('handles empty key_findings', () => {
+  it('handles empty key_findings', () => {
     const input: Record<string, unknown> = {
       score_matrix: [],
     }
@@ -295,7 +295,7 @@ describe('normalizeComparativeOutput', () => {
     expect(result.key_findings).toEqual([])
   })
 
-  test('converts pivot-table format: { participant: { criterion: score } }', () => {
+  it('converts pivot-table format: { participant: { criterion: score } }', () => {
     const input: Record<string, unknown> = {
       bare: { correctness: 4, correctness_rationale: 'works', efficiency: 3, efficiency_rationale: 'ok' },
       tdd: { correctness: 5, correctness_rationale: 'tests', efficiency: 4, efficiency_rationale: 'clean' },
@@ -310,7 +310,7 @@ describe('normalizeComparativeOutput', () => {
     expect(bareCorrectness.weight).toBe(0.25)
   })
 
-  test('clamps score to 1-5 range', () => {
+  it('clamps score to 1-5 range', () => {
     const input: Record<string, unknown> = {
       score_matrix: [
         { participant_id: 'bare', criterion: 'a', weight: 0.5, score: 0, rationale: 'terrible' },
@@ -367,7 +367,7 @@ const manifestWithRubrics: ArenaManifestType = {
 }
 
 describe('buildComparativePrompt with structured criteria', () => {
-  test('injects rubric anchors into prompt', () => {
+  it('injects rubric anchors into prompt', () => {
     const prompt = buildComparativePrompt({ manifest: manifestWithRubrics, verdicts: [] })
     expect(prompt).toContain('信息准确性')
     expect(prompt).toContain('Evaluator: ISTJ测试员')
@@ -377,7 +377,7 @@ describe('buildComparativePrompt with structured criteria', () => {
     expect(prompt).toContain('Evaluator: INTJ架构师')
   })
 
-  test('falls back to bare format for string criteria', () => {
+  it('falls back to bare format for string criteria', () => {
     const manifest: ArenaManifestType = {
       id: 'test', created_at: '2026-01-01T00:00:00Z', task: 'test', mode: 'decks',
       participants: [{ id: 'a', name: 'A', deck: 'd1' }, { id: 'b', name: 'B', deck: 'd2' }],
@@ -393,7 +393,7 @@ describe('buildComparativePrompt with structured criteria', () => {
 // Simulate a realistic LLM judge output — the kind of JSON an actual Claude
 // comparative judge call would produce. Verify our normalization handles it.
 describe('full pipeline: mock LLM output → schema validation', () => {
-  test('clean score_matrix passes through ComparativeReport.parse', () => {
+  it('clean score_matrix passes through ComparativeReport.parse', () => {
     const cleanOutput = {
       score_matrix: [
         { participant_id: 'bare', criterion: 'accuracy', weight: 0.4, score: 3, rationale: 'Correct on Bun version but missed pnpm migration detail' },
@@ -419,7 +419,7 @@ describe('full pipeline: mock LLM output → schema validation', () => {
     expect(deepAccuracy!.score).toBe(5)
   })
 
-  test('messy LLM output with field name variants gets normalized', () => {
+  it('messy LLM output with field name variants gets normalized', () => {
     // Simulates a messy Claude output — participantId instead of participant_id,
     // reason instead of rationale, string score
     const messyLLMOutput = {

package/src/player.test.ts

@@ -1,4 +1,4 @@
-import { describe, test, expect } from 'bun:test'
+import { describe, it, expect } from 'bun:test'
 import { resolvePlayer, resolveSides, groupBySide, totalRuns } from './player'
 import { parseArenaToml } from './arena-toml'
 
@@ -20,29 +20,29 @@ deck = "./decks/rich.toml"
 `)
 
 describe('resolvePlayer', () => {
-  test('maps claude-code → claude', () => {
+  it('maps claude-code → claude', () => {
     expect(resolvePlayer('claude-code')).toBe('claude-sdk')
   })
 
-  test('maps Claude → claude (case insensitive)', () => {
+  it('maps Claude → claude (case insensitive)', () => {
     expect(resolvePlayer('Claude')).toBe('claude-sdk')
   })
 
-  test('maps kimi → kimi', () => {
+  it('maps kimi → kimi', () => {
     expect(resolvePlayer('kimi')).toBe('kimi')
   })
 
-  test('passes through unknown player names', () => {
+  it('passes through unknown player names', () => {
     expect(resolvePlayer('expert-architect')).toBe('expert-architect')
   })
 
-  test('trims whitespace', () => {
+  it('trims whitespace', () => {
     expect(resolvePlayer('  claude-code  ')).toBe('claude-sdk')
   })
 })
 
 describe('resolveSides', () => {
-  test('resolves all sides in arena.toml', () => {
+  it('resolves all sides in arena.toml', () => {
     const sides = resolveSides(toml)
     expect(sides).toHaveLength(2)
     expect(sides[0].platform).toBe('claude-sdk')
@@ -50,7 +50,7 @@ describe('resolveSides', () => {
     expect(sides[0].playerName).toBe('claude-code')
   })
 
-  test('preserves side config', () => {
+  it('preserves side config', () => {
     const sides = resolveSides(toml)
     expect(sides[0].side.name).toBe('minimal')
     expect(sides[0].side.deck).toBe('./decks/minimal.toml')
@@ -58,7 +58,7 @@ describe('resolveSides', () => {
 })
 
 describe('groupBySide', () => {
-  test('groups by side name with run count', () => {
+  it('groups by side name with run count', () => {
     const groups = groupBySide(toml)
     expect(groups).toHaveLength(2)
     expect(groups[0].runs).toBe(3) // runs_per_side
@@ -66,7 +66,7 @@ describe('groupBySide', () => {
     expect(groups[0].platform).toBe('claude-sdk')
   })
 
-  test('control flag preserved', () => {
+  it('control flag preserved', () => {
     const controlToml = parseArenaToml(`
 [arena]
 task = "x"
@@ -89,7 +89,7 @@ control = true
 })
 
 describe('totalRuns', () => {
-  test('calculates sides × runs_per_side', () => {
+  it('calculates sides × runs_per_side', () => {
     expect(totalRuns(toml)).toBe(6) // 2 sides × 3 runs
   })
 })

package/src/preflight.test.ts

@@ -10,7 +10,7 @@
  *   formatSkillWarnings  — warning string generation
  */
 
-import { describe, test, expect } from 'bun:test'
+import { describe, it, expect } from 'bun:test'
 import {
   parseDeckSkills,
   checkSkillExistence,
@@ -26,15 +26,15 @@ import {
 
 describe('parseDeckSkills', () => {
 
-  test('empty deck → empty array', () => {
+  it('empty deck → empty array', () => {
     expect(parseDeckSkills({})).toEqual([])
   })
 
-  test('deck with no skill sections → empty array', () => {
+  it('deck with no skill sections → empty array', () => {
     expect(parseDeckSkills({ deck: { max_cards: 10 } })).toEqual([])
   })
 
-  test('inline-table format: single tool skill with path', () => {
+  it('inline-table format: single tool skill with path', () => {
     const parsed = {
       tool: {
         skills: {
@@ -47,7 +47,7 @@ describe('parseDeckSkills', () => {
     ])
   })
 
-  test('inline-table format: multiple skills', () => {
+  it('inline-table format: multiple skills', () => {
     const parsed = {
       tool: {
         skills: {
@@ -62,7 +62,7 @@ describe('parseDeckSkills', () => {
     ])
   })
 
-  test('array format: skills = ["a", "b"]', () => {
+  it('array format: skills = ["a", "b"]', () => {
     const parsed = {
       tool: {
         skills: ['web-search', 'docx']
@@ -74,7 +74,7 @@ describe('parseDeckSkills', () => {
     ])
   })
 
-  test('innate section parsed separately', () => {
+  it('innate section parsed separately', () => {
     const parsed = {
       innate: {
         skills: {
@@ -93,7 +93,7 @@ describe('parseDeckSkills', () => {
     ])
   })
 
-  test('transient section parsed', () => {
+  it('transient section parsed', () => {
     const parsed = {
       transient: {
         skills: {
@@ -106,7 +106,7 @@ describe('parseDeckSkills', () => {
     ])
   })
 
-  test('object entry without path → path=null', () => {
+  it('object entry without path → path=null', () => {
     const parsed = {
       tool: {
         skills: {
@@ -119,7 +119,7 @@ describe('parseDeckSkills', () => {
     ])
   })
 
-  test('object entry with non-string path → path=null', () => {
+  it('object entry with non-string path → path=null', () => {
     const parsed = {
       tool: {
         skills: {
@@ -132,7 +132,7 @@ describe('parseDeckSkills', () => {
     ])
   })
 
-  test('array entry that is not a string → skipped', () => {
+  it('array entry that is not a string → skipped', () => {
     const parsed = {
       tool: { skills: ['valid', 123, null, 'also-valid'] }
     }
@@ -142,7 +142,7 @@ describe('parseDeckSkills', () => {
     ])
   })
 
-  test('all three sections populated → ordered innate, tool, transient', () => {
+  it('all three sections populated → ordered innate, tool, transient', () => {
     const parsed = {
       innate: { skills: { a: { path: '/a' } } },
       tool: { skills: { b: { path: '/b' } } },
@@ -162,12 +162,12 @@ describe('parseDeckSkills', () => {
 
 describe('checkSkillExistence', () => {
 
-  test('empty skills → empty array', () => {
+  it('empty skills → empty array', () => {
     const exists = (_: string) => true
     expect(checkSkillExistence([], '/cold', exists)).toEqual([])
   })
 
-  test('skill with explicit path → resolves <coldPool>/<path>/SKILL.md', () => {
+  it('skill with explicit path → resolves <coldPool>/<path>/SKILL.md', () => {
     const exists = (p: string) => p === '/cold/github.com/owner/repo/skills/my-skill/SKILL.md'
     const skills = [{ name: 'my-skill', path: 'github.com/owner/repo/skills/my-skill', section: 'tool' }]
     const result = checkSkillExistence(skills, '/cold', exists)
@@ -176,7 +176,7 @@ describe('checkSkillExistence', () => {
     ])
   })
 
-  test('skill without path (array format) → resolves <coldPool>/<name>/SKILL.md', () => {
+  it('skill without path (array format) → resolves <coldPool>/<name>/SKILL.md', () => {
     const exists = (p: string) => p === '/cold/web-search/SKILL.md'
     const skills = [{ name: 'web-search', path: null, section: 'tool' }]
     const result = checkSkillExistence(skills, '/cold', exists)
@@ -185,7 +185,7 @@ describe('checkSkillExistence', () => {
     ])
   })
 
-  test('HTTP path → uses name as fallback for path resolution', () => {
+  it('HTTP path → uses name as fallback for path resolution', () => {
     const exists = (p: string) => p === '/cold/my-skill/SKILL.md'
     const skills = [{ name: 'my-skill', path: 'https://example.com/deck.toml', section: 'tool' }]
     const result = checkSkillExistence(skills, '/cold', exists)
@@ -194,7 +194,7 @@ describe('checkSkillExistence', () => {
     ])
   })
 
-  test('all found → all found=true', () => {
+  it('all found → all found=true', () => {
     const exists = (_: string) => true
     const skills = [
       { name: 'a', path: '/a', section: 'tool' },
@@ -206,7 +206,7 @@ describe('checkSkillExistence', () => {
     ])
   })
 
-  test('some missing → mixed found/not-found', () => {
+  it('some missing → mixed found/not-found', () => {
     const exists = (p: string) => p.includes('a')
     const skills = [
       { name: 'a', path: '/a', section: 'tool' },
@@ -218,7 +218,7 @@ describe('checkSkillExistence', () => {
     ])
   })
 
-  test('different coldPoolDir → different expectedPath prefix', () => {
+  it('different coldPoolDir → different expectedPath prefix', () => {
     const exists = (_: string) => true
     const skills = [{ name: 'x', path: 'github.com/x', section: 'tool' }]
     const a = checkSkillExistence(skills, '/home/user/.agents/skill-repos', exists)
@@ -234,35 +234,35 @@ describe('checkSkillExistence', () => {
 
 describe('validateLinkResult', () => {
 
-  test('exitCode 0 → ok', () => {
+  it('exitCode 0 → ok', () => {
     expect(validateLinkResult(0, '')).toEqual({ ok: true })
   })
 
-  test('exitCode 0 with stderr → still ok (stderr is not always errors)', () => {
+  it('exitCode 0 with stderr → still ok (stderr is not always errors)', () => {
     expect(validateLinkResult(0, 'some warning output')).toEqual({ ok: true })
   })
 
-  test('exitCode 1 → not ok, error contains snippet', () => {
+  it('exitCode 1 → not ok, error contains snippet', () => {
     const result = validateLinkResult(1, 'something went wrong')
     expect(result.ok).toBe(false)
     expect(result.error).toContain('exit 1')
     expect(result.error).toContain('something went wrong')
   })
 
-  test('exitCode null → not ok (null !== 0)', () => {
+  it('exitCode null → not ok (null !== 0)', () => {
     const result = validateLinkResult(null, 'process killed')
     expect(result.ok).toBe(false)
     expect(result.error).toContain('exit null')
   })
 
-  test('stderr truncated to 300 chars in error message', () => {
+  it('stderr truncated to 300 chars in error message', () => {
     const longStderr = 'x'.repeat(500)
     const result = validateLinkResult(1, longStderr)
     expect(result.ok).toBe(false)
     expect(result.error!.length).toBeLessThan(350) // "Deck link failed (exit 1): " + 300 chars
   })
 
-  test('exitCode 0, empty stderr → ok with no error field', () => {
+  it('exitCode 0, empty stderr → ok with no error field', () => {
     const result = validateLinkResult(0, '')
     expect(result.ok).toBe(true)
     expect(result.error).toBeUndefined()
@@ -275,16 +275,16 @@ describe('validateLinkResult', () => {
 
 describe('buildCopyPlan', () => {
 
-  test('empty entries → empty plan', () => {
+  it('empty entries → empty plan', () => {
     expect(buildCopyPlan('/work', '/out', [], new Set())).toEqual([])
   })
 
-  test('all skipped → empty plan', () => {
+  it('all skipped → empty plan', () => {
     const skip = new Set(['.claude', 'skill-deck.toml'])
     expect(buildCopyPlan('/work', '/out', ['.claude', 'skill-deck.toml'], skip)).toEqual([])
   })
 
-  test('normal entries → mapped to outDir', () => {
+  it('normal entries → mapped to outDir', () => {
     const skip = new Set<string>()
     expect(buildCopyPlan('/work', '/out', ['output.md', 'report.docx'], skip)).toEqual([
       { src: '/work/output.md', dest: '/out/output.md', name: 'output.md' },
@@ -292,7 +292,7 @@ describe('buildCopyPlan', () => {
     ])
   })
 
-  test('mixed skip and non-skip → only non-skipped', () => {
+  it('mixed skip and non-skip → only non-skipped', () => {
     const skip = new Set(['.claude', 'skill-deck.toml', 'skill-deck.lock'])
     const entries = ['.claude', 'output.md', 'skill-deck.toml', 'report.docx', 'skill-deck.lock']
     expect(buildCopyPlan('/work', '/out', entries, skip)).toEqual([
@@ -301,13 +301,13 @@ describe('buildCopyPlan', () => {
     ])
   })
 
-  test('preserves entry order', () => {
+  it('preserves entry order', () => {
     const skip = new Set<string>()
     const entries = ['c', 'a', 'b']
     expect(buildCopyPlan('/w', '/o', entries, skip).map(e => e.name)).toEqual(['c', 'a', 'b'])
   })
 
-  test('nested paths work (agent-produced subdirectories)', () => {
+  it('nested paths work (agent-produced subdirectories)', () => {
     const skip = new Set<string>()
     expect(buildCopyPlan('/work', '/out', ['subdir/output.pdf'], skip)).toEqual([
       { src: '/work/subdir/output.pdf', dest: '/out/subdir/output.pdf', name: 'subdir/output.pdf' },
@@ -321,30 +321,30 @@ describe('buildCopyPlan', () => {
 
 describe('resolveColdPoolDir', () => {
 
-  test('explicit absolute path → returned as-is', () => {
+  it('explicit absolute path → returned as-is', () => {
     expect(resolveColdPoolDir('/opt/cold', '/home/user', '/fallback')).toBe('/opt/cold')
   })
 
-  test('explicit relative path → returned as-is', () => {
+  it('explicit relative path → returned as-is', () => {
     expect(resolveColdPoolDir('my-cold-pool', '/home/user', '/fallback')).toBe('my-cold-pool')
   })
 
-  test('tilde path → expanded with homeDir', () => {
+  it('tilde path → expanded with homeDir', () => {
     expect(resolveColdPoolDir('~/.agents/skill-repos', '/home/user', '/fallback'))
       .toBe('/home/user/.agents/skill-repos')
   })
 
-  test('tilde at start only → expanded; tilde elsewhere not expanded', () => {
+  it('tilde at start only → expanded; tilde elsewhere not expanded', () => {
     expect(resolveColdPoolDir('path/with~/tilde', '/home/user', '/fallback'))
       .toBe('path/with~/tilde')
   })
 
-  test('undefined → uses fallback', () => {
+  it('undefined → uses fallback', () => {
     expect(resolveColdPoolDir(undefined, '/home/user', '/default/cold'))
       .toBe('/default/cold')
   })
 
-  test('empty string → uses fallback (|| operator)', () => {
+  it('empty string → uses fallback (|| operator)', () => {
     expect(resolveColdPoolDir('', '/home/user', '/default/cold'))
       .toBe('/default/cold')
   })
@@ -356,7 +356,7 @@ describe('resolveColdPoolDir', () => {
 
 describe('formatSkillWarnings', () => {
 
-  test('all found → empty array', () => {
+  it('all found → empty array', () => {
     const checks = [
       { name: 'a', expectedPath: '/p/a/SKILL.md', found: true, section: 'tool' },
       { name: 'b', expectedPath: '/p/b/SKILL.md', found: true, section: 'tool' },
@@ -364,7 +364,7 @@ describe('formatSkillWarnings', () => {
     expect(formatSkillWarnings(checks)).toEqual([])
   })
 
-  test('some missing → one warning per missing skill', () => {
+  it('some missing → one warning per missing skill', () => {
     const checks = [
       { name: 'pdf', expectedPath: '/cold/pdf/SKILL.md', found: false, section: 'tool' },
       { name: 'docx', expectedPath: '/cold/docx/SKILL.md', found: true, section: 'tool' },
@@ -374,7 +374,7 @@ describe('formatSkillWarnings', () => {
     ])
   })
 
-  test('all missing → warning for each', () => {
+  it('all missing → warning for each', () => {
     const checks = [
       { name: 'a', expectedPath: '/p/a/SKILL.md', found: false, section: 'innate' },
       { name: 'b', expectedPath: '/p/b/SKILL.md', found: false, section: 'tool' },
@@ -382,11 +382,11 @@ describe('formatSkillWarnings', () => {
     expect(formatSkillWarnings(checks)).toHaveLength(2)
   })
 
-  test('empty array → empty array', () => {
+  it('empty array → empty array', () => {
     expect(formatSkillWarnings([])).toEqual([])
   })
 
-  test('section name appears in warning string', () => {
+  it('section name appears in warning string', () => {
     const checks = [
       { name: 'x', expectedPath: '/p/x', found: false, section: 'transient' },
     ]
@@ -411,14 +411,14 @@ const TMP = '/tmp/arena-test'
 
 describe('buildArchiveSidePlan', () => {
 
-  test('default: sides=["."] maps to fromDir', () => {
+  it('default: sides=["."] maps to fromDir', () => {
     const plan = buildArchiveSidePlan(TMP, ['.'], _p => true)
     expect(plan).toEqual([
       { side: '.', sourceDir: TMP, found: true },
     ])
   })
 
-  test('single side, subdirectory exists → source = fromDir/side', () => {
+  it('single side, subdirectory exists → source = fromDir/side', () => {
     const exists = (p: string) => p === pathJoin(TMP, 'side-a')
     const plan = buildArchiveSidePlan(TMP, ['side-a'], exists)
     expect(plan).toEqual([
@@ -426,14 +426,14 @@ describe('buildArchiveSidePlan', () => {
     ])
   })
 
-  test('single side, subdirectory MISSING → fallback to fromDir root', () => {
+  it('single side, subdirectory MISSING → fallback to fromDir root', () => {
     const plan = buildArchiveSidePlan(TMP, ['side-a'], _p => false)
     expect(plan).toEqual([
       { side: 'side-a', sourceDir: TMP, found: true },
     ])
   })
 
-  test('multi side, all subdirectories exist', () => {
+  it('multi side, all subdirectories exist', () => {
     const exists = (p: string) =>
       p === pathJoin(TMP, 'side-a') || p === pathJoin(TMP, 'side-b')
     const plan = buildArchiveSidePlan(TMP, ['side-a', 'side-b'], exists)
@@ -443,7 +443,7 @@ describe('buildArchiveSidePlan', () => {
     ])
   })
 
-  test('multi side, one missing → found=false (caller handles warn+skip)', () => {
+  it('multi side, one missing → found=false (caller handles warn+skip)', () => {
     const exists = (p: string) => p === pathJoin(TMP, 'side-a')
     const plan = buildArchiveSidePlan(TMP, ['side-a', 'side-b'], exists)
     expect(plan).toEqual([
@@ -452,19 +452,19 @@ describe('buildArchiveSidePlan', () => {
     ])
   })
 
-  test('"." side does NOT trigger fallback when missing (found=false)', () => {
+  it('"." side does NOT trigger fallback when missing (found=false)', () => {
     const plan = buildArchiveSidePlan(TMP, ['.'], _p => false)
     expect(plan).toEqual([
       { side: '.', sourceDir: TMP, found: false },
     ])
   })
 
-  test('empty sides array → empty plan', () => {
+  it('empty sides array → empty plan', () => {
     const plan = buildArchiveSidePlan(TMP, [], _p => true)
     expect(plan).toEqual([])
   })
 
-  test('three sides, middle missing', () => {
+  it('three sides, middle missing', () => {
     const exists = (p: string) =>
       p === pathJoin(TMP, 'side-a') || p === pathJoin(TMP, 'side-c')
     const plan = buildArchiveSidePlan(TMP, ['side-a', 'side-b', 'side-c'], exists)
@@ -509,7 +509,7 @@ path = "github.com/anthropics/skills/skills/pdf"
 
 describe('buildPreparePlan', () => {
 
-  test('single skill deck → plan with 1 skill, hasSkills=true', () => {
+  it('single skill deck → plan with 1 skill, hasSkills=true', () => {
     const plan = buildPreparePlan({
       deckPath: '/tmp/test-deck.toml',
       deckContent: DECK_ONE_SKILL,
@@ -524,7 +524,7 @@ describe('buildPreparePlan', () => {
     expect(plan.deckPath).toBe('/tmp/test-deck.toml')
   })
 
-  test('empty deck → skills=[], hasSkills=false', () => {
+  it('empty deck → skills=[], hasSkills=false', () => {
     const plan = buildPreparePlan({
       deckPath: '/tmp/empty.toml',
       deckContent: DECK_EMPTY,
@@ -535,7 +535,7 @@ describe('buildPreparePlan', () => {
     expect(plan.hasSkills).toBe(false)
   })
 
-  test('two skills (innate + tool) → both parsed with correct sections', () => {
+  it('two skills (innate + tool) → both parsed with correct sections', () => {
     const plan = buildPreparePlan({
       deckPath: '/tmp/two.toml',
       deckContent: DECK_TWO_SKILLS,
@@ -548,7 +548,7 @@ describe('buildPreparePlan', () => {
     expect(plan.hasSkills).toBe(true)
   })
 
-  test('AGENTS.md contains mandatory sections', () => {
+  it('AGENTS.md contains mandatory sections', () => {
     const plan = buildPreparePlan({
       deckPath: '/tmp/d.toml',
       deckContent: DECK_ONE_SKILL,
@@ -561,7 +561,7 @@ describe('buildPreparePlan', () => {
     expect(plan.agentsMd).toContain('skill-deck.toml')
   })
 
-  test('invalid TOML → skills=[], hasSkills=false (no crash)', () => {
+  it('invalid TOML → skills=[], hasSkills=false (no crash)', () => {
     const plan = buildPreparePlan({
       deckPath: '/tmp/bad.toml',
       deckContent: 'this is not toml {{{',
@@ -572,7 +572,7 @@ describe('buildPreparePlan', () => {
     expect(plan.hasSkills).toBe(false)
   })
 
-  test('deckContent is preserved in plan', () => {
+  it('deckContent is preserved in plan', () => {
     const plan = buildPreparePlan({
       deckPath: '/tmp/d.toml',
       deckContent: DECK_ONE_SKILL,

package/src/runner.test.ts

@@ -1,9 +1,9 @@
-import { describe, test, expect } from 'bun:test'
+import { describe, it, expect } from 'bun:test'
 import { buildArenaPrompt, formatPlanOutput, runArenaFromToml, type ArenaIO } from './runner'
 import { buildExecutionPlan, parseArenaToml } from './arena-toml'
 
 describe('buildArenaPrompt — plan-mode (pure, no IO)', () => {
-  test('includes CWD, Deck, and output directory', () => {
+  it('includes CWD, Deck, and output directory', () => {
     const prompt = buildArenaPrompt({
       brief: 'Write a hello world function',
       cwd: '/tmp/arena-cell',
@@ -14,7 +14,7 @@ describe('buildArenaPrompt — plan-mode (pure, no IO)', () => {
     expect(prompt).toContain('Produce output to: /tmp/arena-cell/')
   })
 
-  test('includes decision-log.jsonl mandatory instructions', () => {
+  it('includes decision-log.jsonl mandatory instructions', () => {
     const prompt = buildArenaPrompt({
       brief: 'test',
       cwd: '/tmp/arena-cell',
@@ -27,7 +27,7 @@ describe('buildArenaPrompt — plan-mode (pure, no IO)', () => {
     expect(prompt).toContain('reason (why)')
   })
 
-  test('includes robustness and tools instructions', () => {
+  it('includes robustness and tools instructions', () => {
     const prompt = buildArenaPrompt({
       brief: 'test',
       cwd: '/tmp/arena-cell',
@@ -38,7 +38,7 @@ describe('buildArenaPrompt — plan-mode (pure, no IO)', () => {
     expect(prompt).toContain('.claude/skills/')
   })
 
-  test('task brief appears at the end', () => {
+  it('task brief appears at the end', () => {
     const prompt = buildArenaPrompt({
       brief: 'Generate a dark-mode CSS theme',
       cwd: '/tmp/arena-cell',
@@ -54,7 +54,7 @@ describe('buildArenaPrompt — plan-mode (pure, no IO)', () => {
     expect(briefIdx).toBeGreaterThan(toolsIdx)
   })
 
-  test('outputDir overrides default output path', () => {
+  it('outputDir overrides default output path', () => {
     const prompt = buildArenaPrompt({
       brief: 'test',
       cwd: '/tmp/arena-cell',
@@ -64,7 +64,7 @@ describe('buildArenaPrompt — plan-mode (pure, no IO)', () => {
     expect(prompt).toContain('Produce output to: /custom/output/')
   })
 
-  test('preflightReport included when provided', () => {
+  it('preflightReport included when provided', () => {
     const prompt = buildArenaPrompt({
       brief: 'test',
       cwd: '/tmp/arena-cell',
@@ -75,7 +75,7 @@ describe('buildArenaPrompt — plan-mode (pure, no IO)', () => {
     expect(prompt).toContain('✅ 3 skills linked, 0 missing')
   })
 
-  test('no preflight section when report not provided', () => {
+  it('no preflight section when report not provided', () => {
     const prompt = buildArenaPrompt({
       brief: 'test',
       cwd: '/tmp/arena-cell',
@@ -84,14 +84,14 @@ describe('buildArenaPrompt — plan-mode (pure, no IO)', () => {
     expect(prompt).not.toContain('Preflight:')
   })
 
-  test('prompt is deterministic — same inputs = same output', () => {
+  it('prompt is deterministic — same inputs = same output', () => {
     const opts = { brief: 'test', cwd: '/tmp/a', deckPath: '/tmp/d.toml' }
     expect(buildArenaPrompt(opts)).toBe(buildArenaPrompt(opts))
   })
 })
 
 describe('runArenaFromToml — dry-run mode (no fs/spawn/agent calls)', () => {
-  test('returns plan without executing cells', async () => {
+  it('returns plan without executing cells', async () => {
     const toml = parseArenaToml(`
 [arena]
 task = "Write a hello world function"
@@ -125,7 +125,7 @@ deck = "/tmp/deck.toml"
     expect(logs.some(l => l.includes('Dry-run'))).toBe(true)
   })
 
-  test('dry-run does not call io.mkdir or io.agentSpawn', async () => {
+  it('dry-run does not call io.mkdir or io.agentSpawn', async () => {
     const toml = parseArenaToml(`
 [arena]
 task = "Write a hello world function"
@@ -164,7 +164,7 @@ deck = "/tmp/deck.toml"
 })
 
 describe('runArenaFromToml — single cell execution with mock agentSpawn', () => {
-  test('executes one cell and writes outputs via io', async () => {
+  it('executes one cell and writes outputs via io', async () => {
     const toml = parseArenaToml(`
 [arena]
 task = "Write a hello world function"
@@ -234,7 +234,7 @@ deck = "/tmp/deck.toml"
 })
 
 describe('runArenaFromToml — error recovery path', () => {
-  test('cell exception produces ERROR verdict and continues', async () => {
+  it('cell exception produces ERROR verdict and continues', async () => {
     const toml = parseArenaToml(`
 [arena]
 task = "Write a hello world function"
@@ -296,7 +296,7 @@ deck = "/tmp/deck.toml"
     expect(verdict.reason).toContain('Agent spawn failed')
   })
 
-  test('spawn non-zero exit logs warning but continues', async () => {
+  it('spawn non-zero exit logs warning but continues', async () => {
     const toml = parseArenaToml(`
 [arena]
 task = "Write a hello world function"

package/src/stats.test.ts

@@ -1,4 +1,4 @@
-import { describe, test, expect } from 'bun:test'
+import { describe, it, expect } from 'bun:test'
 import { aggregateSideStats, aggregateAllStats } from './stats'
 import type { JudgeVerdict } from '@lythos/test-utils/schema'
 
@@ -14,7 +14,7 @@ function makeVerdict(overrides?: Partial<JudgeVerdict>): JudgeVerdict {
 // ── aggregateSideStats ─────────────────────────────────────────────────────
 
 describe('aggregateSideStats', () => {
-  test('single run: passRate=1, no variance', () => {
+  it('single run: passRate=1, no variance', () => {
     const stats = aggregateSideStats('test', [makeVerdict()])
     expect(stats.sideName).toBe('test')
     expect(stats.runs).toBe(1)
@@ -23,7 +23,7 @@ describe('aggregateSideStats', () => {
     expect(stats.errorRate).toBe(0)
   })
 
-  test('3 runs: 2 PASS, 1 FAIL', () => {
+  it('3 runs: 2 PASS, 1 FAIL', () => {
     const verdicts = [
       makeVerdict(),
       makeVerdict(),
@@ -34,7 +34,7 @@ describe('aggregateSideStats', () => {
     expect(stats.failRate).toBeCloseTo(1 / 3)
   })
 
-  test('confidence: mean across runs', () => {
+  it('confidence: mean across runs', () => {
     const verdicts = [
       makeVerdict({ confidence: 90 }),
       makeVerdict({ confidence: 80 }),
@@ -45,13 +45,13 @@ describe('aggregateSideStats', () => {
     expect(stats.confidenceVariance).toBeCloseTo(100) // (100+0+100)/2 = 100
   })
 
-  test('confidence: null when no verdict has it', () => {
+  it('confidence: null when no verdict has it', () => {
     const stats = aggregateSideStats('test', [makeVerdict(), makeVerdict()])
     expect(stats.meanConfidence).toBeNull()
     expect(stats.confidenceVariance).toBeNull()
   })
 
-  test('per-criterion pass rate', () => {
+  it('per-criterion pass rate', () => {
     const verdicts = [
       makeVerdict({ criteria: [{ name: 'accuracy', passed: true }] }),
       makeVerdict({ criteria: [{ name: 'accuracy', passed: false }] }),
@@ -63,7 +63,7 @@ describe('aggregateSideStats', () => {
     expect(stats.criteria[0].mean).toBeCloseTo(2 / 3)
   })
 
-  test('per-criterion scores: mean and variance', () => {
+  it('per-criterion scores: mean and variance', () => {
     const verdicts = [
       makeVerdict({ scores: { coverage: 5, relevance: 4 } }),
       makeVerdict({ scores: { coverage: 3, relevance: 4 } }),
@@ -75,14 +75,14 @@ describe('aggregateSideStats', () => {
     expect(stats.scoreByCriterion.relevance.variance).toBe(0) // all 4s
   })
 
-  test('zero runs: all zeros', () => {
+  it('zero runs: all zeros', () => {
     const stats = aggregateSideStats('empty', [])
     expect(stats.runs).toBe(0)
     expect(stats.passRate).toBe(0)
     expect(stats.meanConfidence).toBeNull()
   })
 
-  test('handles ERROR verdicts correctly', () => {
+  it('handles ERROR verdicts correctly', () => {
     const verdicts = [
       makeVerdict(),
       makeVerdict({ verdict: 'ERROR', reason: 'parse failed' }),
@@ -96,7 +96,7 @@ describe('aggregateSideStats', () => {
 // ── aggregateAllStats ──────────────────────────────────────────────────────
 
 describe('aggregateAllStats', () => {
-  test('aggregates multiple sides', () => {
+  it('aggregates multiple sides', () => {
     const map = new Map<string, JudgeVerdict[]>()
     map.set('side-a', [makeVerdict(), makeVerdict()])
     map.set('side-b', [makeVerdict({ verdict: 'FAIL', reason: 'nope' })])