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 d4rkNinja/arcforge --skill ai-agent-system-architecturegit clone --depth 1 https://github.com/d4rkNinja/arcforgeWrote 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/d4rkninja/arcforge/ai-agent-system-architecture)<a href="https://agentmods.dev/skills/d4rkninja/arcforge/ai-agent-system-architecture"><img src="https://agentmods.dev/badge/skills/d4rkninja/arcforge/ai-agent-system-architecture/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/skills/d4rkninja/arcforge/ai-agent-system-architecture"><img src="https://agentmods.dev/badge/skills/d4rkninja/arcforge/ai-agent-system-architecture.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 1 finding, up to high
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- high YARA Match · line 3 YARA rule matched a hack tool or exploit indicator (offensive tools, reconnaissance, privilege escalation, or exploit frameworks).Fix: Remove offensive tool references and exploit code. Legitimate agent skills should not contain penetration testing tools, exploit frameworks, or reconnaissance utilities.
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 | $0.00091 | $0.04880 |
| Opus 5 | $0.00046 | $0.02440 |
| Sonnet 5 | $0.00018 | $0.00976 |
| Haiku 4.5 | $0.00009 | $0.00488 |
Grade A, and why
ai-agent-system-architecture 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 5d 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 — 366 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Think Through AI & Agent Systems
Overview
Design AI products as governed software systems rather than prompts wrapped in an API. Separate the probabilistic model from the deterministic harness that controls context, tools, memory, policy, evaluation, recovery, and irreversible actions.
Core principle: The model proposes; the harness constrains, verifies, records, and decides what may happen next.
NO PRODUCTION AI FLOW WITHOUT:
1. a measurable task and acceptance rubric;
2. explicit model, data, context, memory, and tool boundaries;
3. a failure and fallback path;
4. offline and online evaluation;
5. traceable cost, latency, safety, and version evidence.
When to Use
Use this skill for:
- LLM-backed features and assistants;
- retrieval-augmented generation and document intelligence;
- tool-using agents, coding agents, and autonomous workflows;
- supervisor/worker or multi-agent systems;
- model gateways, routing, fallbacks, and provider abstraction;
- long-running jobs with memory, checkpoints, and resumability;
- AI actions affecting money, permissions, infrastructure, customers, or external systems;
- AI quality, safety, observability, inference cost, and production-readiness reviews.
Do not use model generation where deterministic code, search, rules, or a normal workflow can satisfy the requirement more reliably and cheaply.
Operating Modes
| Mode | Trigger | Primary result |
|---|---|---|
| Think | The AI-system decision is not settled; includes discovery, design, optimization, and migration framing | task contract, baseline, risk class, alternatives, governed design, and validation path |
| Review | Prompts, RAG, tools, agents, or operating evidence already exist | evidence-backed findings, eval gaps, and blockers |
| Change | Decisions are approved and repository changes are requested | bounded control, prompt, retrieval, tool, or rollout changes plus pending proof |
| Verify | Quality, safety, latency, cost, or authority claims need proof | observed evaluation and operational evidence with residual risks |
What ships with it
20 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.
- agents/openai.yaml 263 B
- assets/ai-production-readiness-checklist.md 5.0 KB
- assets/ai-system-spec-template.md 829 B
- assets/evaluation-plan-template.md 621 B
- assets/tool-contract-template.md 687 B
- examples/bounded-agent-platform-example.md 4.3 KB
- examples/governed-support-agent.md 1.2 KB
- references/01-harness-control-plane.md 3.7 KB
- references/02-context-retrieval-memory.md 2.9 KB
- references/03-tool-security-governance.md 2.6 KB
- references/04-evaluation-observability.md 2.9 KB
- references/05-model-serving-economics.md 2.2 KB
- references/06-classical-and-ai-control-obligations.md 6.9 KB
- references/papers/140-ai-llm-backend-fundamentals.md 40 KB
- references/papers/141-agent-execution.md 37 KB
- references/papers/142-ai-memory.md 36 KB
- references/papers/143-rag-infrastructure.md 36 KB
- references/papers/144-untrusted-code-execution.md 37 KB
- references/papers/145-plugin-extension-architecture.md 35 KB
- references/production-ai-patterns.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.
- 5d ago Changed · +2 lines ebe1ddfdee36
- 9d ago First seen · 364 lines · 91 tokens per session scan A d4dceee26993
ai-agent-system-architecture is a skill published in the GitHub repository d4rkNinja/arcforge (16 stars, last pushed 6d ago), licensed MIT. It adds 91 tokens to every session and 4,880 once invoked, about $0.0005 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
portaljs-architect
Recommend a data-portal architecture (storage, compute, catalog, access, hosting, metadata) from stated needs, then hand off to the build skills. The advisory entry point. Use when starting a new data-portal project and the underlying architecture has not yet been decided.
senior-data-engineer
World-class data engineering skill for building scalable data pipelines, ETL/ELT systems, real-time streaming, and data infrastructure. Expertise in Python, SQL, Spark, Airflow, dbt, Kafka, Flink, Kinesis, and modern data stack. Includes data modeling, pipeline orchestration, data quality, streaming quality…
architecture-paradigm-pipeline
Applies pipes-and-filters for sequential data transformations. Use when data flows through discrete stages like ETL, streaming analytics, or CI/CD pipelines.
data-mesh-expert
Expert-level data mesh architecture, domain-oriented ownership, data products, federated governance, and self-serve platforms. Use when the user mentions architecture, domain driven, data products, governance, or platform, or when the task involves Data Mesh Principles, Domain-Oriented Data Ownership, Data as a…
data-data-platform
Use this skill when designing end-to-end data platforms: data lake architecture, data lakehouse architecture, distributed storage (S3, ADLS, GCS), distributed compute (Spark, Trino, Presto), data catalog (Datahub, Amundsen, Marquez), data mesh, data versioning (LakeFS, DVC), data virtualization (Dremio, Starburst).…
nosql-database-design
Designs a NoSQL data model by leading with access pattern analysis. Covers DynamoDB single-table design (PK/SK/GSI) and MongoDB embedding vs referencing, consistency models, and capacity planning. Invoked when the user asks to design a DynamoDB schema, MongoDB data model, or NoSQL data model.