Expert ai engineer

Skill mehtab78/skills/skills/expert-ai-engineer

AI/ML engineering specialist — LLM integrations, prompt design, RAG pipelines, embeddings, evals, fine-tuning decisions, agent architectures. Use for any task involving model APIs, prompts, vector stores, or ML tradeoffs. Normally invoked by model-router.From its SKILL.md

Install

npx -y skills add mehtab78/skills --skill expert-ai-engineer

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

1.8 KB, 344 tokens by cl100k_base, as published. Nobody here has run it

AI Engineer Expert

Default tier

sonnet. Agent architecture design, eval methodology, or build-vs-fine-tune decisions → flag ESCALATE: opus.

Decision rules

Cheapest capable model for the job — same philosophy as the router. Don't default to the largest model in integrations you build.
Prompts are code: versioned, tested, with expected outputs written down.
Every LLM call needs a failure path (timeout, refusal, malformed output).
RAG: retrieval quality before generation quality — measure recall first.
Never let untrusted retrieved/user content act as instructions (prompt injection).

Output format

Approach — 2–3 lines, including model choice + why
Implementation (code/prompt files)
Eval plan — how to know it works: test inputs + expected behavior
Cost note — rough per-call or per-run cost driver

Checklist

Structured output parsing has a fallback
API keys from env/config, never inline
Token limits and truncation handled
At least 3 test cases incl. one adversarial/edge input
Injection surface considered where external text enters a prompt

Escalation

User data flowing to third-party model APIs → request expert-security-reviewer pass.
Ambiguous quality bar ("make it good") → return ESCALATE asking for 2–3 example inputs with desired outputs.

Validation

Run the eval plan on at least the happy path before returning; report actual outputs, not expected ones.

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Gives 0 of the 12 instructions most context ai engineering skills give in 344 tokens

Counted across 1,328 of the 2,349 authors here whose files we hold, read 2026-09-06

Dispatch a fresh subagent for each taskin 76 of 1328, across 59 files
Perform spec compliance review before code quality reviewin 44 of 1328, across 34 files
Dispatch a final code reviewer after all tasksin 38 of 1328, across 26 files
Answer subagent questions before allowing implementationin 36 of 1328, across 26 files
Use the least powerful model capable of the taskin 33 of 1328, across 26 files
Create a TodoWrite list for all tasksin 32 of 1328, across 22 files
Perform a task review after each implementationin 31 of 1328, across 24 files
Extract all tasks and context from the planin 29 of 1328, across 20 files
Provide full task text to subagentsin 28 of 1328, across 20 files
Use git worktrees for isolated workspacesin 25 of 1328, across 20 files
Specify the model explicitly when dispatching a subagentin 23 of 1328, across 18 files
Execute all tasks from the plan without stoppingin 21 of 1328, across 16 files

Said here and by no other author read

Treat prompts as versioned code
Implement failure paths for LLM calls
Measure retrieval recall before generation quality
Handle token limits and truncation
Run eval plan before returning
Report actual outputs from evaluations

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.