Getting reliable work out of coding agents: instruction and context files, task scoping, reviewing agent-written code, repository conventions, hooks and workflow guardrails.
ls /playbooks --problems=10
Playbooks
Recommendations for working on AI projects, collected from practitioners' write-ups and grouped by problem. Every item links to the article and the original source.
recommendations 0
Structuring agents: orchestration and multi-agent patterns, tool and MCP server design, memory, planning loops, state and handoffs.
Sandboxing, scoped credentials, prompt-injection defence, permission and approval gates, egress control, supply chain of models and tools.
Grounding and checking what a model claims: citation and quote verification, hallucination detection, LLM-as-judge pitfalls, structured-output validation, human approval of actions.
Measuring and regression-testing LLM apps and agents: eval datasets, benchmarks you run yourself, CI gates, tests on tool calls and traces.
Chunking, embeddings, keyword and hybrid search, reranking, vector stores, filtering, and measuring retrieval quality.
Cutting token spend and API bills: model routing, prompt and KV caching, batching, context budgets, moving work from models to code.
Running models yourself: GPU memory, quantization, KV cache, serving engines, throughput and latency tuning, hardware choices.
Monitoring, tracing and alerting for LLM apps, incident response, rollouts, model upgrades and provider migrations, reliability.
Fine-tuning, distillation, RL and preference tuning, synthetic data, dataset curation and labeling for models you train.