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 soden46/engineer-flow --skill ai-llm-engineeringgit clone --depth 1 https://github.com/soden46/engineer-flowWrote 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/soden46/engineer-flow/ai-llm-engineering)<a href="https://agentmods.dev/skills/soden46/engineer-flow/ai-llm-engineering"><img src="https://agentmods.dev/badge/skills/soden46/engineer-flow/ai-llm-engineering.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 5.1 | $0.00031 | $0.00314 |
| Opus 5 | $0.00015 | $0.00157 |
| Sonnet 5 | $0.00006 | $0.00063 |
| Haiku 4.5 | $0.00003 | $0.00031 |
Grade A, and why
ai-llm-engineering 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 7d 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.
What it actually says
ai-llm-engineering
Use this skill for LLM integrations, agents, prompts, embeddings, retrieval, evaluation, model workflows, and AI pipelines.
Principles
Treat model output as untrusted and nondeterministic.
Separate:
- model instructions
- application logic
- tools
- retrieval
- persistence
- evaluation
Define expected outputs and failure behavior.
For structured output use enforceable schemas where available.
For tool use:
- validate arguments
- enforce authorization outside the model
- limit tool capability
- verify side effects
For retrieval systems evaluate both retrieval quality and final answer quality.
For prompts:
- state the task clearly
- provide relevant context
- avoid irrelevant context
- define output constraints where useful
For evaluation use representative cases and frozen test sets when comparing changes.
Do not tune against held-out evaluation cases.
For expensive model workloads consider:
- latency
- token usage
- caching
- batching
- retries
- rate limits
- fallback behavior
Do not treat model confidence as proof of correctness.
Adaptation
Use project evidence to determine the actual language, framework, runtime, and existing conventions.
When stack-specific implementation guidance is needed, prefer project evidence, native framework or language mechanisms, and relevant user-installed specialist skills. Technology-specific guidance must not redefine or weaken the core engineering requirement.
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.
- 7d ago First seen · 65 lines · 31 tokens per session scan A d483ed9a69ae
ai-llm-engineering is a skill published in the GitHub repository soden46/engineer-flow (3 stars, last pushed 4d ago), licensed MIT. It adds 31 tokens to every session and 314 once invoked, about $0.0002 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 2026-08-31.
Other skills, from other repositories
promptly-prompt
Clarity gate for prompts: intercepts unclear requests and forces the model to restate the user's intent to align before working, then diagnose the root cause and reuse the domain's mature, established practices before improvising. Clear prompts pass through untouched.
hunt-llm-ai
Hunt LLM/AI feature bugs — prompt injection, indirect injection, exfiltration via tool-use, ASCII smuggling, agentic AI security framework (ASI01-ASI10). Patterns: direct prompt injection in user input (bypass system prompt with 'ignore previous instructions'), indirect injection via documents/web pages the model…
enhance-prompt
Transforms vague UI ideas into polished, Stitch-optimized prompts. Enhances specificity, adds UI/UX keywords, injects design system context, and structures output for better generation results.
implementing-llm-guardrails-for-security
Implements input and output validation guardrails for LLM-powered applications to prevent prompt injection, data leakage, toxic content generation, and hallucinated outputs. Builds a security validation pipeline using NVIDIA NeMo Guardrails Colang definitions, custom Python validators for PII detection and content…
ai-prompt-writing
Write effective prompts for AI coding tools — structure task context, specify constraints, provide examples, and iterate based on output quality. Targets developers using Claude Code, Copilot, Cursor, and similar tools.
prompt-engineering-guide
Design, test, and optimize LLM prompts systematically — with evaluation frameworks, chain-of-thought patterns, output formatting, and iteration methodology for reliable AI outputs.