Explore all skills by this author →

Verified:GitHub

agentic-engineering

Automate AI engineering workflows with eval-first loops and tiered model routing.

1|Updated Mar 23, 2026

Search Tags:#workflow #decomposition #ai-agent #risk-management #model-routing #agentic-engineering #eval-first

Ready to run

npx skills add https://github.com/sbalagan22/bloomr --skill agentic-engineering-sbalagan22

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: agentic-engineering
Source: https://github.com/sbalagan22/bloomr/tree/main/.claude/skills/agentic-engineering
Command: npx skills add https://github.com/sbalagan22/bloomr --skill agentic-engineering-sbalagan22

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Enables AI-driven engineering by enforcing eval-first execution, safe decomposition, and cost-aware routing to produce verifiable implementations.

Core Features & Use Cases

Eval-First Loop: Define capability eval, run baseline, execute, and re-evaluate deltas for safety and quality.
Task Decomposition: Break work into independently verifiable units with single dominant risk.
Model Routing: Route tasks by complexity using tiered models for efficiency and safety.
Session Strategy: Maintain continuity across related units and clear progression after milestones.
Review Focus for AI-Generated Code: Emphasize invariants, error boundaries, security, and risk.
Cost Discipline: Track model usage, tokens, retries, and wall-clock time to optimize cost.

Quick Start

Provide a high-level goal and let the system decompose into agent-sized units, route execution by complexity, and perform eval-first checks.

Frequently Asked Questions about agentic-engineering

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

What is eval-first execution in AI-driven engineering workflows?▼

It is a development loop that defines capability evaluations, runs a baseline, executes the task, and re-evaluates deltas to ensure safety and quality in AI-driven engineering workflows.

How do I decompose complex engineering tasks for AI agents?▼

Break work into independently verifiable units, ensuring each unit carries a single dominant risk for safer automated execution.

How does model routing reduce costs for AI workflows?▼

It routes tasks to tiered models based on complexity, tracking token usage, retries, and wall-clock time to maintain auditable cost discipline.

Can I track token usage and wall-clock time for AI-generated code?▼

Yes, the system tracks model usage, tokens, retries, and wall-clock time to enforce cost discipline and maintain auditable cost tracking.

What should I review when using AI-generated code in production?▼

Review invariants, error boundaries, security, and risk factors to ensure the implementation meets rigorous safety standards.

Does this approach maintain session continuity across related task units?▼

Yes, the session strategy maintains continuity across related units while clearing progression after milestones to ensure reproducible outputs.

samymity