Agentic Engineering
If you're letting Claude write most of your code, this formalizes how to stay in control without micromanaging. It's built around eval-first execution (write the test, capture baseline failures, implement, check deltas), decomposition into 15-minute verifiable units, and cost-aware routing between Haiku, Sonnet, and Opus based on task complexity. The review guidance is sharp: focus on invariants, error boundaries, and hidden coupling, not style bikeshedding. Useful if you're already doing agentic workflows and want clearer discipline around when to start fresh sessions, when to escalate model tiers, and what actually matters in code review when the agent did the typing.
- skills.sh installs
- 5.8K
- GitHub stars
- 300.3K
- last updated
- Aug 21, 2026
Install
$ npx -y skills add affaan-m/ecc --skill agentic-engineering --agent claude-codeInstalls into .claude/skills of the current project.
Files
Agentic Engineering
Use this skill for engineering workflows where AI agents perform most implementation work and humans enforce quality and risk controls.
Operating Principles
- Define completion criteria before execution.
- Decompose work into agent-sized units.
- Route model tiers by task complexity.
- Measure with evals and regression checks.
Eval-First Loop
- Define capability eval and regression eval.
- Run baseline and capture failure signatures.
- Execute implementation.
- Re-run evals and compare deltas.
Task Decomposition
Apply the 15-minute unit rule:
- each unit should be independently verifiable
- each unit should have a single dominant risk
- each unit should expose a clear done condition
Model Routing
- Haiku: classification, boilerplate transforms, narrow edits
- Sonnet: implementation and refactors
- Opus: architecture, root-cause analysis, multi-file invariants
Session Strategy
- Continue session for closely-coupled units.
- Start fresh session after major phase transitions.
- Compact after milestone completion, not during active debugging.
Review Focus for AI-Generated Code
Sponsors
Agent CaffeineSponsoredClose your lid, come back to finished work
SnugSponsoredThe clipboard manager for the AI eraPromote your product hereReach 125k+ visitors monthly
WaxSponsoredBlock anything with supervision, without erasing your iPhonePromote your product hereReach 125k+ visitors monthly
knobsSponsoredSee your site on any phone, any day, in any languageDetails
- Repository
- github.com/affaan-m/ecc
- Category
- AI & Agent Building
- First seen
- Jul 14, 2026
More from affaan-m/ecc
All 263- Vite Patterns5.8K
- Bun Runtime5.8K
- Agent Architecture Audit5.8K
- Nodejs Keccak2565.8K
- Security Scan5.8K
- Cpp Testing5.8K
- Django Tdd5.8K
- Dashboard Builder5.7K
- Redis Patterns5.7K
- Eval Harness5.7K
- Manim Video5.7K
- Content Engine5.7K
- Springboot Tdd5.7K
- Motion Patterns5.7K
- Configure Ecc5.7K
- Token Budget Advisor5.7K
- Investor Materials5.7K
- Agent Introspection Debugging5.7K
- Laravel Verification5.7K
- Marketing Campaign5.7K
- Motion Foundations5.7K
- Kotlin Testing5.7K
- Ai Regression Testing5.7K
- Dart Flutter Patterns5.7K
Recommended
More AI & Agent Building
review-agent-setupwshobson/agentsConfigure human-in-the-loop gating for AI agent review actions in Claude Code. Use when setting up a project where an agent may post PR reviews, comments, merge5.6Kskills.sh installs
agent-harness-constructionaffaan-m/everything-claude-codeagent harness construction5.5Kskills.sh installsagent-payment-x402affaan-m/eccAdd x402 payment execution to AI agents with per-task budgets, spending controls, and non-custodial wallets. Supports Base through agentwallet-sdk and X Layer t5.5Kskills.sh installscost-aware-llm-pipelineaffaan-m/everything-claude-codecost aware llm pipeline5.4Kskills.sh installs
project-builderstarchild-ai-agent/official-skillsEnd-to-end project engineering: design, incremental build, verify, debug systematically. Use when building software, dashboards, scheduled jobs, or web apps the5.2Kskills.sh installs