mirror of
https://fastgit.cc/github.com/Yeachan-Heo/oh-my-claudecode
synced 2026-04-30 22:01:58 +08:00
3a833c3395
Extend the existing harsh-critic benchmark framework with reusable benchmarks for code-reviewer, debugger, and executor agents. Enables measurable prompt tuning by comparing old (pre-consolidation) vs new (merged) prompts with ground-truth scoring. New infrastructure: - benchmarks/shared/ — generalized scoring types, parser, reporter, runner - benchmarks/code-reviewer/ — 3 fixtures (SQL injection, clean code, payment edge cases) - benchmarks/debugger/ — 3 fixtures (React undefined, Redis intermittent, TS build errors) - benchmarks/executor/ — 3 fixtures (trivial, scoped, complex tasks) - benchmarks/run-all.ts — top-level orchestrator with --save-baseline and --compare modes - npm scripts: bench:prompts, bench:prompts:save, bench:prompts:compare Each benchmark includes archived pre-consolidation prompts for reproducible comparison even after old agent files are deleted. Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
58 lines
2.8 KiB
JSON
58 lines
2.8 KiB
JSON
{
|
|
"fixtureId": "task-notification-refactor",
|
|
"fixturePath": "fixtures/tasks/task-notification-refactor.md",
|
|
"domain": "task",
|
|
"expectedVerdict": "complex",
|
|
"isCleanBaseline": false,
|
|
"findings": [
|
|
{
|
|
"id": "IMPL-NR-1",
|
|
"severity": "CRITICAL",
|
|
"category": "finding",
|
|
"summary": "Must define a NotificationChannel interface with a send method for the strategy pattern",
|
|
"keywords": ["NotificationChannel", "interface", "send", "strategy", "pattern"],
|
|
"explanation": "The core abstraction is a NotificationChannel interface with a send(notification) method. This enables the strategy pattern for channel routing."
|
|
},
|
|
{
|
|
"id": "IMPL-NR-2",
|
|
"severity": "CRITICAL",
|
|
"category": "finding",
|
|
"summary": "Must implement EmailChannel, SmsChannel, and PushChannel classes",
|
|
"keywords": ["EmailChannel", "SmsChannel", "PushChannel", "class", "implement"],
|
|
"explanation": "Three concrete channel implementations are required. Each should handle its own sending logic, error handling, and status tracking."
|
|
},
|
|
{
|
|
"id": "IMPL-NR-3",
|
|
"severity": "CRITICAL",
|
|
"category": "finding",
|
|
"summary": "Must maintain backward compatibility — existing callers without channels param default to email",
|
|
"keywords": ["backward", "compatibility", "default", "email", "existing"],
|
|
"explanation": "Existing code calls sendNotification without a channels parameter. The refactored version must default to email channel to avoid breaking existing callers."
|
|
},
|
|
{
|
|
"id": "IMPL-NR-4",
|
|
"severity": "MAJOR",
|
|
"category": "finding",
|
|
"summary": "Should route notifications based on user preferences — lookup preferences per user",
|
|
"keywords": ["user", "preferences", "route", "channel", "lookup"],
|
|
"explanation": "The NotificationService should look up user preferences to determine which channels to use. Users with SMS enabled should receive SMS notifications, etc."
|
|
},
|
|
{
|
|
"id": "IMPL-NR-5",
|
|
"severity": "MAJOR",
|
|
"category": "finding",
|
|
"summary": "Each channel should independently track status and handle errors — one channel failure shouldn't block others",
|
|
"keywords": ["independent", "status", "error", "failure", "block", "channel"],
|
|
"explanation": "If email sending fails, SMS and push should still be attempted. Each channel independently records its status (sent/failed) in the database."
|
|
},
|
|
{
|
|
"id": "IMPL-NR-6",
|
|
"severity": "MINOR",
|
|
"category": "finding",
|
|
"summary": "API route should accept optional channels override parameter",
|
|
"keywords": ["API", "route", "channels", "override", "parameter", "optional"],
|
|
"explanation": "The POST /notifications endpoint should accept an optional channels array to override user preferences for specific notifications."
|
|
}
|
|
]
|
|
}
|