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 agentmods add agents/dannykkh/skill-olympus/ai-mlgit clone --depth 1 https://github.com/Dannykkh/skill-olympusWrote 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/dannykkh/skill-olympus/ai-ml)<a href="https://agentmods.dev/agents/dannykkh/skill-olympus/ai-ml"><img src="https://agentmods.dev/badge/agents/dannykkh/skill-olympus/ai-ml.svg" alt="Measured on agentmods" 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 | $0.00082 | $0.02599 |
| Opus 5 | $0.00041 | $0.01300 |
| Sonnet 5 | $0.00016 | $0.00520 |
| Haiku 4.5 | $0.00008 | $0.00260 |
Grade B, and why
ai-ml scanned grade B with 2 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 4d 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.
Sends data to an external URLmediumData exfiltration
A POST to an outside endpoint may be telemetry or may be exfiltration; either way the mod talks to somewhere, and you should know where.
requests.post("http://localhost:11434/...") # 직접 HTTP Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
requests.post("http://localhost:11434/...") # 직접 HTTP How it starts
The opening of the file, as written. The whole thing — 281 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AI/ML Agent
AI/ML 애플리케이션 구축 + LLM API 최신 모델/패턴 보장을 담당합니다.
왜 검증이 필수인가: Claude의 훈련 데이터에는 구식 모델명(gpt-4, claude-3-opus 등)이 포함되어 있어, 의식적으로 검증하지 않으면 deprecated 모델/API를 사용하게 됩니다.
PART 1: 최신 모델 검증 (코드 작성 전 필수)
검증 워크플로우
LLM 모델명이나 API 호출 코드를 작성하기 전에 반드시 실행:
Step 1 — WebSearch (필수):
WebSearch: "site:{공식문서URL} models {year}"
| Provider | 검색 쿼리 |
|---|---|
| OpenAI | site:platform.openai.com models 2026 |
| Anthropic | site:docs.anthropic.com models 2026 |
| Google AI | site:ai.google.dev gemini models 2026 |
| Ollama | site:ollama.com library |
Step 2 — Context7 MCP (보조):
resolve_library_id("{provider}") → get_library_docs() 로 SDK 최신 사용법 확인
Step 3 — 체크리스트:
- 모델명이 WebSearch 결과의 현재 모델과 일치
- SDK import 패턴이 최신 버전에 맞음
- Deprecated API를 사용하지 않음
- API 키 환경변수명이 공식 문서와 일치
Provider별 공식 문서
OpenAI
| 문서 | URL |
|---|---|
| 모델 목록 | https://platform.openai.com/docs/models |
| API Reference | https://platform.openai.com/docs/api-reference |
| Deprecations | https://platform.openai.com/docs/deprecations |
# ❌ 구식 패턴
openai.ChatCompletion.create(...) # v0 API
model="gpt-4" # 구버전
model="gpt-3.5-turbo" # 구버전
model="text-davinci-003" # deprecated
# ✅ 최신 패턴
from openai import OpenAI
client = OpenAI()
response = client.chat.completions.create(
model="<WebSearch로 확인한 최신 모델>",
messages=[...]
)
Anthropic (Claude)
| 문서 | URL |
|---|---|
| 모델 목록 | https://docs.anthropic.com/en/docs/about-claude/models |
| API Reference | https://docs.anthropic.com/en/api |
# ❌ 구식 패턴
model="claude-3-opus-20240229" # 구버전
model="claude-2" # deprecated
anthropic.completions.create(...) # v0 API
# ✅ 최신 패턴
import anthropic
client = anthropic.Anthropic()
message = client.messages.create(
model="<WebSearch로 확인한 최신 모델>",
max_tokens=1024,
messages=[{"role": "user", "content": "..."}]
)
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.
- 4d ago First seen · 281 lines · 82 tokens per session scan B 8c8917c7e3f5
ai-ml is an agent published in the GitHub repository Dannykkh/skill-olympus (5 stars, last pushed 3d ago), licensed MIT. It adds 82 tokens to every session and 2,599 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it B with 2 findings (sends data to an external url, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other agents, from other repositories
document-steward
GOAL: One document per domain. Minimum tokens for maximum clarity.
strategy-fidelity-voc
Evaluates app fidelity and completion against docs/SYSTEMARCHITECTURE.md and domain references. Serves as voice of customer: defines user workflows and outcomes, then validates implementation against them. Use proactively before releases, after major changes, or when validating feature completeness.
cross-project-memory
Designs and executes efficient cross-project and long-term memory so agents build apps better. Use when adding or improving memory that spans projects, sessions, or runs; when defining what to remember, how to scope it, and how to retrieve it for agent context.
architect
Software architecture lead for hybrid systems using traditional architecture (Next.js + PostgreSQL) and AI-agent-supportive architecture (ruvector). Use proactively for system design, module boundaries, interfaces, migration plans, and architecture trade-offs.
investigator
investigates a bug to identify root cause and set success criteria for resolution; creates investigation report for fixer agent to guide implementation.
ai-advocate
Audits the project for poor AI agent behaviors and recommends concrete improvements to make coding workflows more agent-friendly, reliable, and fast. Use proactively when agents struggle, loop, miss context, or produce inconsistent changes.