Agentic Engineering
SkillAI & modelsOnce added, this skill lets your AI take on engineering work on its own: it breaks a big task into smaller steps, runs evaluations to steer and check its work, and picks models with cost in mind. The result is an AI that can plan, carry out, and verify engineering tasks rather than just answering questions about them.
Available today. Use it from your connected AI after setup.
No other account needed.
After adding the skill, give your AI an engineering task and ask it to break the work into steps before starting. While it works, you can ask it to show the evaluations it ran and how it picked models for each step.
Then ask your AI: use the Agentic Engineering skill
What your AI can do with it
- Break large engineering tasks into smaller, manageable steps
- Run evaluations as a first step to guide and verify its work
- Choose between models based on the cost of each task
- Carry out multi-step engineering work on its own
- Use evaluation results to decide when a task is done
What this skill tells your AI
The instructions your AI receives, as published by contextgo/contextgo in src/process/resources/assistant/engineering/everything-in-claude-code/skills/agentic-engineering/SKILL.md and read by ahel’s review.
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
Prioritize:
- invariants and edge cases
- error boundaries
- security and auth assumptions
- hidden coupling and rollout risk
Do not waste review cycles on style-only disagreements when automated format/lint already enforce style.
Cost Discipline
Track per task:
- model
- token estimate
- retries
- wall-clock time
- success/failure
Escalate model tier only when lower tier fails with a clear reasoning gap.
Signals
- GitHub stars
- 54
- Forks
- 5
- Last commit
- May 2026
Advanced
- Catalog kind
- skill
- Gateway key
agentic-engineering-contextgo- Source
- github.com/contextgo/contextgo