senior-prompt-engineer

これはレビュー時に保存された抜粋です。完全で最新の内容は外部ソースを確認してください。

---
name: "senior-prompt-engineer"
description: Use when the user asks to optimize prompts, design prompt templates, evaluate LLM outputs with an eval set, measure RAG retrieval quality, validate agent/tool configurations, analyze token usage, or design structured-output contracts. Covers eval-driven prompt iteration, RAG metrics (relevance, faithfulness, coverage), agent workflow validation, and token/cost budgeting — all model-agnostic, with three stdlib Python tools.
---

# Senior Prompt Engineer

Eval-driven prompt engineering, RAG quality measurement, and agent workflow validation. Everything here is **model-agnostic by design**: techniques are framed by what they do, not by which model generation they were observed on, and the tools never hardcode model IDs or pricing — you supply your provider's current rates when you want dollar figures.

## Operating Rules

1. **Never change a prompt without a baseline.** Capture metrics first (`--analyze --output baseline.json`), then compare every iteration against it.
2. **Eval set before optimization.** 10–20 representative cases with expected outputs minimum. If the user has no eval set, build one with them before touching the prompt — optimizing against vibes is the #1 failure mode.
3. **Prefer platform features over prompt hacks.** If the provider offers native structured outputs / JSON schema enforcement, tool-use APIs, or prompt caching, use those instead of "respond ONLY with JSON" incantations. Prompt-level format enforcement is the fallback, not the default.
4. **Current-generation models need less scaffolding.** Don't add chain-of-thought boilerplate, role framing, or few-shot examples reflexively — frontier models often do worse with redundant scaffolding. Add each element only when the eval set shows it helps.
5. **Cost numbers are always user-supplied.** Look up the provider's current per-Mtok pricing and pass it via `--price-per-mtok` (never trust a cached price table — including any you remember).

## Tools (exact CLIs, all stdlib)

### 1. Prompt Optimizer — `scripts/prompt_optimizer.py`

Static analysis: token estimate, clarity/structure scores (0–100), ambiguity + redundancy detection, few-shot example extraction.

```bash
# Full analysis (human-readable report)
python3 scripts/prompt_optimizer.py prompt.txt --analyze

# Save machine-readable baseline for later comparison
python3 scripts/prompt_optimizer.py prompt.txt --analyze --json --output baseline.json

# Token estimate; cost only if you supply your provider's current rate
python3 scripts/prompt_optimizer.py prompt.txt --tokens --model claude --price-per-mtok 3.00

# Whitespace/redundancy-trimmed version
python3 scripts/prompt_optimizer.py prompt.txt --optimize --output optimized.txt

# Extract Input/Output few-shot pairs to JSON
python3 scripts/prompt_optimizer.py prompt.txt --extract-examples --output examples.json

# Compare a revision against the saved baseline
python3 scripts/prompt_optimizer.py optimized.txt --analyze --compare baseline.json
```

`--model` accepts any string; only the tokenizer family is inferred (names containing "claude" → 3.5 chars/token, otherwise 4.0). Exit 0 on success, 1 on missing file.

### 2. RAG Evaluator — `scripts/rag_evaluator.py`

Measures retrieval and grounding quality from two JSON files (formats printed in `--help`).

```bash
python3 scripts/rag_evaluator.py --contexts retrieved.json --questions eval_set.json
python3 scripts/rag_evaluator.py --contexts ctx.json --questions q.json --k 10 --json
python3 scripts/rag_evaluator.py --contexts ctx.json --questions q.json --output report.json --verbose
python3 scripts/rag_evaluator.py --contexts ctx.json --questions q.json --compare baseline_report.json
```

Reports context relevance, precision@k, coverage, answer faithfulness, groundedness. Treat relevance < 0.80 as a retrieval problem (chunking/embedding/filtering), not a prompt problem — fix retrieval before rewriting the generation prompt.

### 3. Agent Orchestrator — `scr