بنقرة واحدة
principles
Guiding principles and decision-making patterns for the ikigai project
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
القائمة
Guiding principles and decision-making patterns for the ikigai project
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
استنادا إلى تصنيف SOC المهني
Automated quality check loops with escalation and fix sub-agents
JSON-based end-to-end test format, runner, and mock provider
Jujutsu (jj) skill for the ikigai project
How to write effective Ralph goals for Ikigai-driven workflows
Create and manage Ralph goals from Ikigai using the real ralph-pipeline scripts
Create repositories using the real ralph-pipeline repo-create script
| name | principles |
| description | Guiding principles and decision-making patterns for the ikigai project |
Foundational beliefs that guide recommendations. When proposing solutions, align with these principles.
Correctness is non-negotiable. 90% test coverage, every branch, no exceptions. Quality gates exist because willpower fails - systems enforce standards.
Explicit over implicit. Ownership is visible. Errors are typed. Naming is precise. Decisions are documented. If it's not obvious, make it obvious.
Simplicity through discipline, not cleverness. Single-threaded. Hierarchical memory. Crash on impossible states. Complexity is a cost - pay it only when forced.
When uncertain, fail loudly. PANIC over silent corruption. Assert liberally. Unknown states are unacceptable.
Choose battle-tested foundations. C, PostgreSQL, talloc, direct terminal rendering. Solid platforms over trendy tooling.
Split rather than sprawl. 16KB file limits. One module, one responsibility. Refactor before it hurts.
Validate at boundaries, trust internal code. External input is suspect. Internal invariants are asserted. No defensive coding inside trust boundaries.
Spend generously preparing, save ruthlessly executing. Research phase is thorough. Task files are complete. Execution loops are mechanical and cheap.
Reject: feature flags for hypothetical futures, backwards-compat hacks, "hard to test" excuses, implicit ownership, documentation as afterthought, clever code, over-engineering.