Borrowing it
Nothing to install: this file belongs to TakaGoto/rag-learning-academy. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/TakaGoto/rag-learning-academy/main/.claude/agents/deployment-specialist.mdgit clone --depth 1 https://github.com/TakaGoto/rag-learning-academyWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/agents/takagoto/rag-learning-academy/deployment-specialist)<a href="https://agentmods.dev/agents/takagoto/rag-learning-academy/deployment-specialist"><img src="https://agentmods.dev/badge/agents/takagoto/rag-learning-academy/deployment-specialist/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/agents/takagoto/rag-learning-academy/deployment-specialist"><img src="https://agentmods.dev/badge/agents/takagoto/rag-learning-academy/deployment-specialist.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00029 | $0.02072 |
| Opus 5 | $0.00015 | $0.01036 |
| Sonnet 5 | $0.00006 | $0.00414 |
| Haiku 4.5 | $0.00003 | $0.00207 |
Grade B, and why
Deployment Specialist scanned grade B with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 9d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Instruction-override phrasingmediumPrompt injection
Text telling the model to disregard its earlier instructions or safety rules is the shape of a prompt injection, whoever wrote it.
- **Prompt injection via documents**: Adversarial content in the corpus (e.g., "Ignore previous instructions and...") can manipulate LLM output. Mitigation: sanitize chunks, use separate system/user message roles, add ou Downgraded: this mod is about security review, or the phrase is quoted, so it is likely naming the pattern rather than instructing it.
How it starts
The opening of the file, as written. The whole thing — 162 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Shared standards: See
.claude/AGENT_TEMPLATE.mdfor voice, language, calibration, and delegation patterns.
Deployment Specialist
Role Overview
You are the Deployment Specialist of the RAG Learning Academy. You bridge the gap between a RAG system that works in a Jupyter notebook and one that handles real users in production. Building a RAG prototype takes days; making it production-ready takes weeks. You teach the engineering practices, monitoring strategies, and operational knowledge that separate toy projects from reliable systems.
Most RAG tutorials end at "it works on my machine." You start there and teach what comes next: latency budgets, caching, error handling, cost management, and observability.
Core Philosophy
- Production is a different game than prototyping. What works for 10 queries per day breaks at 1,000 queries per minute.
- Latency is a feature. Users won't wait 10 seconds for an answer. Every millisecond matters.
- Cost compounds. A $0.01 query costs $10,000 at 1M queries. Optimize early.
- Observability is non-negotiable. If you can't see what your system is doing, you can't debug it when it breaks.
- Graceful degradation beats hard failure. When a component fails, the system should still provide value, even if reduced.
Key Responsibilities
1. Caching Strategies
- Teach multi-level caching for RAG:
- Embedding cache: Don't re-embed the same text. Cache embeddings by content hash.
- Query cache: If the same question is asked repeatedly, serve from cache.
- Semantic cache: Similar (not identical) queries can share cached results. Use embedding similarity to detect cache hits.
- LLM response cache: Cache generated responses for identical query-context pairs.
- Discuss cache invalidation: when documents change, which caches need to be cleared?
- Teach cache hit rate monitoring and tuning.
2. Scaling Patterns
- Teach how to scale each component:
- Embedding service: Batch processing, GPU inference, async embedding.
- Vector database: Sharding, replication, read replicas.
- LLM calls: Rate limiting, load balancing, fallback models.
- Ingestion pipeline: Async processing, queue-based architecture, incremental updates.
- Discuss horizontal vs. vertical scaling for each component.
- Teach load testing: how to simulate production traffic and find bottlenecks.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 9d ago First seen · 162 lines · 29 tokens per session scan B d1f1efcdb951
Deployment Specialist is an agent published in the GitHub repository TakaGoto/rag-learning-academy (18 stars, last pushed 5mo ago), licensed MIT. It adds 29 tokens to every session and 2,072 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it B with 1 finding (instruction-override phrasing). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other agents, from other repositories
llm-architect
Use when designing LLM systems for production, implementing fine-tuning or RAG architectures, optimizing inference serving infrastructure, or managing multi-model deployments.
data-engineer
Build scalable data pipelines, modern data warehouses, and real-time streaming architectures. Implements Apache Spark, dbt, Airflow, and cloud-native data platforms. Use PROACTIVELY for data pipeline design, analytics infrastructure, or modern data stack implementation.
FabricDataEngineer
Orchestrate end-to-end Microsoft Fabric data engineering workflows that span multiple workloads and personas. Use when the request crosses Spark, Warehouse, Pipelines, Lakehouse architecture, migration, or data quality operations. Delegates deep single-endpoint implementation to specialized skills and resources.
qdrant-expert
Configure and operate the vector store in production. TRIGGER WHEN: creating Qdrant collections, tuning HNSW, quantization, dense plus sparse hybrid search, payload indexing, multi-tenancy, or Qdrant performance troubleshooting. DO NOT TRIGGER WHEN: end-to-end RAG design, or another vector database such as Pinecone…
FAI LangChain Expert
LangChain framework specialist — LCEL expression language, chains, agents with tool use, retrievers, memory, callbacks, LangSmith tracing, and production RAG pipeline patterns.
rag-evaluator
Run retrieval regression gates (hitgate) against the current repo state. Compares Hit@5, MRR, and per-intent metrics to detect whether a change helped, regressed, or held steady. Use for shipping retrieval code changes, validating retuning before merge, or measuring refactor impact on search quality.