342건 · 관련도순
Adds and tunes a reranker stage for RAG using the retrieve-wide, rerank-narrow pattern. Use when relevant documents appear in the initial top-N but are buried by noise, top-1 precision or MRR is low despite adequate recall, or too many marginal chunks consume context. Covers cross-encoder, API, and local rerankers; N-to-k selection; and latency/cost tradeoffs. Do not use when retrieval recall itself is failing.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
Decision SOP for serving LLMs with vLLM. Covers PagedAttention mental model, quantization/parallelism/batching tradeoffs, OOM triage, and when NOT to use vLLM. Activates when a coder-agent is choosing or tuning an inference engine, debugging vLLM throughput/latency/OOM, or comparing vLLM against TGI/SGLang/TensorRT-LLM/llama.cpp.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
Tool skill — the *first move* in any LM-debugging session: dump the actual rendered prompt the framework sent to the model, before changing anything else. Activate when an LM call produced an unexpected output (wrong answer, schema violation, refusal, truncation, cost spike, latency spike, infinite loop, "model got dumber after upgrade"). The skill enforces a 30-second inspect step BEFORE any prompt edit, model swap, retry, or temperature tweak. Cross-framework cheat sheet: DSPy `inspect_history`, LangGraph `get_state_history`, CrewAI `step_callback`, LangChain `set_debug`/`set_verbose`, Aider `/diff`+`--verbose`, raw OpenAI/Anthropic via `OPENAI_LOG=debug`/`ANTHROPIC_LOG=debug` or HTTPX event hooks. Do NOT activate for first-time prompt authoring, exploratory prompt design, or non-LM bugs.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
Comprehensive best practices for robot perception systems covering cameras, LiDARs, depth sensors, IMUs, and multi-sensor setups. Use this skill when working with RGB image processing, depth maps, point clouds, sensor calibration (intrinsic, extrinsic, hand-eye), object detection, semantic segmentation, 3D reconstruction, visual servoing, or perception pipeline optimization. Trigger whenever the user mentions OpenCV, Open3D, PCL, RealSense, ZED, OAK-D, camera calibration, AprilTags, ArUco markers, stereo vision, RGBD, point cloud filtering, ICP registration, coordinate transforms, camera intrinsics, distortion correction, image undistortion, sensor streaming, frame synchronization, or any computer vision task in a robotics context. Also covers multi-camera rigs, time synchronization across sensors, perception latency budgets, and production deployment of perception pipelines.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
Use when evaluating a newly adapted model's accuracy/eval score. Runs evalscope eval (or the test/srt/run_eval.py fallback) against an OpenAI-compatible endpoint on datasets like gsm8k / aime, producing per-dataset reports and a score. Accuracy only — throughput/latency benchmarks belong to model-speed-benchmark. Triggers on accuracy, eval, evalscope, gsm8k, aime, run_eval — not throughput.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
Load tests API endpoints with progressive concurrency. Measures response times, error rates, throughput, and identifies breaking points. Generates a detailed report with latency percentiles, throughput curves, bottleneck analysis, and optimization recommendations.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
The on-call triage procedure for checkout-api. Use whenever you are asked to investigate elevated errors, elevated latency, or a suspected regression.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
Use when adding Spring AI-specific model observations, token usage, latency, externally configured cost attribution, advisor telemetry, or protected prompt and completion logging. Use production-observability for general service metrics, health, logs, and OTLP setup.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
Audit local AI coding-agent sessions with agenttrace for cost, tool failures, latency, anomalies, health, diffs, and CI gates.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
Instrument AI agents with tracing, token metrics, latency, and cost visibility. Use for reliability and debugging.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
Plan, diagnose, and verify performance budgets, latency targets, load tests, capacity estimates, bottleneck analysis, caching strategy, query efficiency, queue throughput, and regression gates. Use when a feature may be slow, a system must scale, a performance regression is suspected, or release readiness depends on throughput, cost, memory, CPU, or response time.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
Applies the engineering and research philosophies of Jeff Dean, Chief Scientist at Google DeepMind and Google Research. Reach for this skill whenever you are designing large-scale distributed systems, optimizing latency and energy efficiency, or making architectural decisions about machine learning infrastructure. It should trigger automatically for topics involving hardware-ML co-design, model distillation, sparse activation, massively multi-task models, or scaling systems by 5x to 10x. Use this skill to evaluate system bottlenecks, transition from specialized to unified models, and optimize experimental velocity. Apply his mental models to avoid premature 100x scaling and to treat AI models as reasoning engines rather than memorization databases.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
Optimize slow PostgreSQL inserts in Python using psycopg2. Use when: (1) Row-by-row inserts are taking too long over network, (2) executemany() isn't providing speedup, (3) Migrating large datasets to PostgreSQL, (4) Network latency making individual INSERT statements impractical. The key is using execute_values() from psycopg2.extras instead of executemany() or individual execute() calls.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
Deploy production recommendation systems with feature stores, caching, A/B testing. Use for personalization APIs, low latency serving, or encountering cache invalidation, experiment tracking, quality monitoring issues.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
Deploy ML models with FastAPI, Docker, Kubernetes. Use for serving predictions, containerization, monitoring, drift detection, or encountering latency issues, health check failures, version conflicts.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
Cloudflare Workers performance optimization with CPU, memory, caching, bundle size. Use for slow workers, high latency, cold starts, or encountering CPU limits, memory issues, timeout errors.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
Profile and debug Hermes sessions from their JSONL transcripts. Find a session and its subagents, build a queryable event table, summarize the work as a hierarchical table of contents, break down wall time, inference, tools, tokens, and estimated cost per agent, identify errors and improvement opportunities, and export a shareable Perfetto trace. Use when a user wants to inspect what a Hermes session did, debug agent or subagent activity, understand session cost or latency, create a session timeline, generate a trace, or learn how to improve the next agent run.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
Design privacy-aware observability for AI agents using traces, spans, structured events, metrics, cost attribution, dashboards, alerts, and investigation workflows. Use when instrumenting an agent, debugging intermittent tool or model failures, defining service-level objectives, analyzing latency or spend, auditing agent decisions, or preparing production monitoring.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
Think and work like an expert Embedded Systems Engineer. Use when a task calls for Embedded Systems Engineer judgment. Reasons from hardware timing, interrupt latency, memory maps, and deterministic WCET resource bounds through DWT CYCCNT and Saleae logic-analyzer timing proofs, SWD/JTAG (J-Link/OpenOCD) with CFSR/HFSR/BFAR fault decode, FreeRTOS/Zephyr respecting configMAX_SYSCALL_INTERRUPT_PRIORITY, MISRA C:2012 and ISO 26262 ASIL discipline, Nordic PPK2 power profiling, and MCUboot signed OTA while treating priority inversion, Cortex-M7 uncached-DMA D-cache incoherence, CAN bus-off, I2C NACK storms, and brownout-during-flash-write NVS corruption as first-class failure modes.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
Think and work like an expert Distributed Systems Researcher. Use when a task calls for Distributed Systems Researcher judgment. Reasons from failure models, consistency contracts, and tail-latency-and-recovery performance through TLA+ model checking, Jepsen and Porcupine linearizability checking, and YCSB/DeathStarBench benchmarking with iptables and kill -9 fault injection, while treating unbounded leases and split-brain, clock skew under NTP, and GC-induced p99 spikes as first-class failure modes.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
Think and work like an expert Database Systems Researcher. Use when a task calls for Database Systems Researcher judgment. Reasons from storage hierarchy, concurrency semantics, query-optimization theory, and declared workload models through TPC-C/H, YCSB, and JOB benchmarks, Jepsen/Elle correctness checkers, and perf/blktrace/fio profiling while treating cardinality-estimation plan regressions, tail-latency spikes under skew, unfair fsync-disabled speedups, and benchmark-trick wins as first-class failure modes.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
Real-world Active Directory environment constraints that silently break attacks when ignored: NTLM disabled (Kerberos fallback), AES-only KDCs (RC4 blocked by GPO), LDAP signing and channel binding requiring LDAPS on 636, Kerberos clock skew (KRB_AP_ERR_SKEW), SPNs that must be FQDNs never short names or IPs (the misleading invalidCredentials), Protected Users / LAPS / gMSA / MachineAccountQuota, LDAP 1000-object paging, and VPN latency. Use this whenever a bind, TGT request, LDAP query, or SMB negotiation fails in a way that looks like bad credentials but is really hardening, or before running any AD tooling against a domain you have not fingerprinted. Standard-tooling knowledge (netexec, impacket, certipy, bloodyAD, kerbrute), no vendor-specific engine.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
Plan and scope production work with Odyssey's real-time interactive video world models (odyssey.ml — Odyssey-1/Odyssey-2/Starchild-1/Agora-1). Use when a request involves generating video that a user steers live with keyboard, text, speech, or controller input (streamed frames that respond in the moment), or when someone must decide between Odyssey-style interactive generation and offline clip generators (Sora/Veo/Kling) or downloadable 3D / game engines. Covers what is verified versus announced, current access routes, latency and coherence limits, feasible-now versus speculative use cases, pricing/access terms, and content/rights considerations. This is a fast-moving research-stage product; the skill teaches honest scoping, not a promise of production reliability.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.
Use Resemble AI Chatterbox for text-to-speech and voice-cloning workflows, including local open-weight Chatterbox, Chatterbox Multilingual, Chatterbox Turbo, and Resemble-hosted Chatterbox API routes. Apply when an agent must choose a Chatterbox variant, prepare reference-voice inputs, control emotion/paralinguistic delivery, plan multilingual TTS, handle PerTh watermarking, or make production decisions about consent, privacy, licensing, latency, and audio QA.
한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.