Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add AnastasiyaW/codex-claude-code-config --skill system-and-data-designgit clone --depth 1 https://github.com/AnastasiyaW/codex-claude-code-configWrote 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/skills/anastasiyaw/codex-claude-code-config/system-and-data-design)<a href="https://agentmods.dev/skills/anastasiyaw/codex-claude-code-config/system-and-data-design"><img src="https://agentmods.dev/badge/skills/anastasiyaw/codex-claude-code-config/system-and-data-design.svg" alt="Measured on agentmods" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00217 | $0.01541 |
| Opus 5 | $0.00109 | $0.00771 |
| Sonnet 5 | $0.00043 | $0.00308 |
| Haiku 4.5 | $0.00022 | $0.00154 |
Grade A, and why
system-and-data-design scanned grade A with 0 findings 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 8d 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.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 123 lines — stays where its author put it; the contents beside it link to each section on GitHub.
System and data design — will it hold, and where does the data live
Two questions that are usually asked together and answered separately, badly. Capacity without storage internals gives a diagram that cannot be built; storage internals without capacity gives a database choice with no reason behind it.
Scope guard — read first
The most common failure here is answering at the wrong scale.
| Situation | Honest answer |
|---|---|
| Internal tool, tens of users | One process, one database, no cache. Stop. |
| Product with real traffic, single region | Estimate first; add a cache and a queue only where the numbers say |
| Multi-region, or data outgrowing one machine | Full pass: estimation → building blocks → replication/partitioning → consistency |
Adding a cache before measuring is the canonical way to convert one problem into two.
Step 1 — requirements before architecture
- Functional: what must it actually do? Write it as verbs, not components.
- Non-functional, with numbers: users, requests/sec at peak, payload size, growth, read:write ratio, acceptable latency, acceptable staleness, retention.
- Constraints: budget, team size, existing stack, compliance, where data may live.
"Acceptable staleness" is the single most useful number and the one nobody asks for. It decides caching, replication and consistency all at once.
Step 2 — back-of-the-envelope, before any diagram
Estimate storage/day, bandwidth at peak, QPS, and working-set size. The point is not precision; it is discovering that the answer is "one machine" or "this cannot work as described" before drawing anything.
Anchors worth remembering: memory reads are ~100ns, SSD ~100µs, a same-region round trip ~0.5ms, cross-continent ~150ms. Anything crossing a network is ~1000× a memory access — which is why one N+1 query pattern outweighs most micro-optimisation.
Step 3 — building blocks, each with a reason
| Block | Add it when | Cost you accept |
|---|---|---|
| Cache | Read-heavy, tolerable staleness, measured hot set | Invalidation becomes your problem |
| Queue | Work can be async; spikes must be absorbed | Ordering, retries, duplicate delivery |
| Read replica | Reads dominate; stale reads acceptable | Replication lag becomes visible to users |
| Partitioning | One machine cannot hold data or throughput | Cross-partition queries and transactions get hard |
| CDN | Static or cacheable content, geographically spread | Purge and versioning discipline |
What ships with it
15 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- references/ddia-systems-original.md 20 KB
- references/ddia-systems/batch-stream.md 14 KB
- references/ddia-systems/data-models.md 12 KB
- references/ddia-systems/fault-tolerance.md 15 KB
- references/ddia-systems/partitioning.md 11 KB
- references/ddia-systems/replication.md 11 KB
- references/ddia-systems/storage-engines.md 11 KB
- references/ddia-systems/transactions.md 13 KB
- references/system-design-original.md 18 KB
- references/system-design/building-blocks.md 11 KB
- references/system-design/common-designs.md 12 KB
- references/system-design/database-scaling.md 11 KB
- references/system-design/estimation-numbers.md 8.4 KB
- references/system-design/four-step-process.md 9.9 KB
- references/system-design/reliability-operations.md 14 KB
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.
- 8d ago First seen · 123 lines · 217 tokens per session scan A 3ea8de05a745
system-and-data-design is a skill published in the GitHub repository AnastasiyaW/codex-claude-code-config (149 stars, last pushed yesterday), licensed MIT. It adds 217 tokens to every session and 1,541 once invoked, about $0.0011 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
stripe-projects
Use when the user wants to provision infrastructure or third-party services using Stripe Projects. Triggers: "I need a database", "set up auth", "add caching", "give me a Postgres", "provision Redis", "I need hosting", "add a vector DB", "get me an API key for X", "get credentials for X", "sign up for a service", "set…
openloomi-api
OpenLoomi ships a local-first HTTP API served from the desktop app (port 3414, fallback 3515). All auth, Memory, AI, RAG, Loop, and Audit data live in a local SQLite database — your data stays on your machine and the OpenLoomi app is the source of truth. The only externally-routed auth path is the Composio OAuth…
firebase-data-connect
Use when setting up Data Connect, writing GraphQL queries/mutations, configuring generated SDKs, handling offline, or applying security rules.
output-dev-credentials
Store and reference encrypted secrets in Output SDK workflows using @outputai/credentials. Use when integrating API keys, database passwords, or third-party tokens.
calling-llms
Use when sending chat completions through liter-llm and routing to a specific provider via the provider/model prefix. Covers the chat call shape, provider routing, modelhint, message roles, and error categories.
running-the-proxy
Use when running the liter-llm api OpenAI-compatible gateway — virtual keys, per-key rate limits, budgets, cost tracking, and model routing. Covers the TOML config and the 22 REST endpoints.