335 skills · sorted by relevance
Tune moflo's memory stack for speed, RAM, and index quality. Covers HNSW parameters (M, efConstruction, ef), vector quantization, batch operations, and common bottlenecks. Use when scaling past ~100k entries or when search latency regresses.
The English summary is being prepared. Showing the original description.
Competitive benchmarks, solution research, and hypothesis generation for identified bottlenecks
The English summary is being prepared. Showing the original description.
Multi-cycle diagnostic pipeline: profile → research → optimize → repeat with full-stack bottleneck analysis
The English summary is being prepared. Showing the original description.
host / 辅助段优化模块（L3）：把**非 DiT、非解码段**的固定开销落成可复验的优化—— ① 交付与搬运（mp4/图片编码、worker→API 通路、落盘、异步化）； ② 装载与预热（权重加载、编译与图下发预热、镜像/容器预热）。 当用户说“mp4 编码/传输/落盘慢”“交付这段占了不少时间”“模型加载慢”“首启编译太久” “固定开销占比高”“每个请求都要重新编码/重新写盘”时使用。 **入口条件**：占比已达门限（瓶颈标签 `非DiT-host段`；标签枚举、门限与测点口径的单一真源在 model-auto-optimization/references/bottleneck-labels.md，本技能只消费不定义）。 near-miss（看似相关但不属本技能）： - **并发 / 吞吐 / 批处理调度 / 队列深度与限流** → **不属本技能**（会改变所有性能口径与 SLO 语义， 未来独立立项；本技能只在**不改变请求并发模型**的前提下压固定开销）； - VAE/TAE 解码段（计算或通信）→ `vae-opt`；DiT 计算 → `dit-perf-opt`； 多卡并行 / 通信掩盖 → `dit-parallel-opt`； - **装库 / 下载权重 / 构建镜像 / 创建容器** → `env-install`（本技能只优化装载与预热**过程本身**）； - 编译与图下发的**机制本体**（pattern / Inductor / NPUGraph / 图 pool / 缓存条目）→ `pattern-dev` / `aclgraph-dev`（本技能不另做一套缓存）； - 产物一致性与精度判据 → `accuracy-gate`；收益口径 → `perf-gate`； - 瓶颈未定位 / 没有占比锚点 → `model-auto-optimization`。
The English summary is being prepared. Showing the original description.
Optimizes application performance. Use when performance requirements exist, when you suspect performance regressions, or when Core Web Vitals or load times need improvement. Use when profiling reveals bottlenecks that need fixing.
The English summary is being prepared. Showing the original description.
Profiles DAG execution performance including latency, token usage, cost, and resource consumption. Identifies bottlenecks and optimization opportunities. Activate on 'performance profile', 'execution metrics', 'latency analysis', 'token usage', 'cost analysis'. NOT for execution tracing (use dag-execution-tracer) or failure analysis (use dag-failure-analyzer).
The English summary is being prepared. Showing the original description.
Diagnose and optimize timing-critical Chisel RTL for FPGA implementation while preserving cycle semantics. Use for Vivado timing bottlenecks, long ready or admission paths, queue and issue selection, free-list bank mapping, priority encoders, one-hot arbitration, wide muxes, late-arriving forwarding or override data, high-fanout controls, cross-module predicates, register-boundary changes, emitted-Verilog inspection, or routed-DCP A/B analysis. Also use when a source-level simplification needs proof that it changes timing without changing architectural state.
The English summary is being prepared. Showing the original description.
Infers person-to-person transmission from pathogen genomes using outbreaker2, TransPhylo, phybreak, BadTrIP, SCOTTI, BEASTLIER, and SNP-distance / cluster-picker approaches (HIV-TRACE for HIV; transcluster). Defines outbreak clusters using pathogen-specific SNP thresholds (NOT a universal cutoff -- TB <=12 SNPs; MRSA <=15; C. difficile <=2; Klebsiella <=21), models within-host diversity and transmission bottlenecks, integrates contact-tracing data, distinguishes generation from serial interval, and attributes source via Bayesian source attribution (islandR). Use when investigating outbreaks for who-infected-whom, defining SNP-cluster outbreak definitions, accounting for unsampled intermediates, choosing between outbreaker2 (rich epi data) and TransPhylo (genomic-only after a dated phylogeny), running source attribution between host populations, calling HIV-TRACE thresholds appropriate to the local subtype, or distinguishing recent transmission from reactivation in TB or chronic HIV.
The English summary is being prepared. Showing the original description.
Quality control for pooled CRISPR screens covering library representation, Gini index, log-skew, replicate Pearson and Spearman concordance, essentialome precision-recall AUC against CEGv2 (Hart 2017), Cas9 cut-toxicity diagnostics, copy-number amplicon detection (Aguirre 2016 / Munoz 2016), bottleneck propagation through plasmid pool, infection, selection, and endpoint stages, MOI verification, and DepMap-style screen-quality scoring. Use when assessing screen quality before hit calling, deciding whether to repeat or rescue a screen, diagnosing low-confidence hits, choosing between MAGeCK / BAGEL2 / Chronos based on quality grade, picking a normalization strategy from QC signatures, or evaluating whether an in-vivo screen retained adequate library complexity.
The English summary is being prepared. Showing the original description.
Queries gnomAD v4 (807k samples), v3, v2.1.1, and constraint metrics with grpmax FAF95, bottleneck-group exclusion, LOEUF interpretation, SV/CNV/mtDNA catalogs, and Whiffin max-credible-AF framework. Use when filtering rare variants, applying ACMG BS1/BA1, ranking genes by LoF intolerance, or selecting between v2 (GRCh37 + chrX/Y constraint) and v4 (GRCh38 + 807k samples).
The English summary is being prepared. Showing the original description.
Optimizes application performance. Use when performance requirements exist, when you suspect performance regressions, or when Core Web Vitals or load times need improvement. Use when profiling reveals bottlenecks that need fixing.
The English summary is being prepared. Showing the original description.
Analyze web performance URLs with the PerfGraph CLI and interpret its deterministic reports, causal chains, and prioritized remediation data. Use when the request names perfgraph, or asks to measure or diagnose loading speed, Core Web Vitals, Lighthouse performance, bottlenecks, or web performance regressions.
The English summary is being prepared. Showing the original description.
[Testing] Use when initializing, writing, reviewing or verifying performance/load/soak/capacity tests. --mode=verify --fix-loop (or --mode=green) diagnoses, fixes and remeasures with performance-review. Static bottleneck audits use performance-review.
The English summary is being prepared. Showing the original description.
Performance audit — bottleneck profiling, N+1 query detection, hot-path analysis; explicit request only, not part of regular feature work.
The English summary is being prepared. Showing the original description.
[Debugging] Use when a workflow step or the user asks for performance bottleneck analysis and validated fixes: queries, memory, network, rendering and concurrency.
The English summary is being prepared. Showing the original description.
Diagnose and fix Kotlin/Native performance bottlenecks with reproducible tests, lowest-level profiling, and minimal expect/actual optimization patches.
The English summary is being prepared. Showing the original description.
Use when performance budgets or SLAs exist, users or monitoring report slowness, Core Web Vitals fall below thresholds, a change may have caused a regression, or a feature handles large datasets or high traffic. Measure-first workflow: profile, fix the proven bottleneck, re-measure. Not for speculative optimization.
The English summary is being prepared. Showing the original description.
Operations leadership for scaling companies. Process design, OKR execution, operational cadence, and scaling playbooks. Use when designing operations, setting up OKRs, building processes, scaling teams, analyzing bottlenecks, planning operational cadence, or when user mentions COO, operations, process improvement, OKRs, scaling, operational efficiency, or execution.
The English summary is being prepared. Showing the original description.
Coaches founders and CEOs on their own development: leadership archetype, delegation, blind spots, and the decisions that change as the company scales. Use when a founder is the bottleneck, struggles to delegate, or is outgrowing their current way of leading.
The English summary is being prepared. Showing the original description.
Use when designing or auditing an agentic AI system holistically. Activates the six-component harness framework (ℛ Reasoning, ℳ Memory, 𝒞 Context, 𝒮 Skills, 𝒪 Orchestration, 𝒢 Governance) from arXiv:2605.26112. Trigger phrases: "design an agent system", "why is my agent unreliable", "harness architecture", "system-level agent design", "beyond model scaling", "agentic infrastructure", "scale my agent system", "agent performance bottleneck".
The English summary is being prepared. Showing the original description.
Review audit-related documents (control design documents, bottleneck analyses, requirements definitions, etc.) for quality, scoring them 0-100 with a severity-rated findings list. Use when reviewing audit documents, checking control design quality, or verifying cross-document consistency. Supports documents governed by US GAAP, IFRS, or J-GAAP.
The English summary is being prepared. Showing the original description.
AI PR intake and multi-agent governance for deployless repositories. Use when a repo has high-volume bot or agent PRs, parallel agent work, noisy low-context PRs, merge-queue pressure, CI exhaustion, duplicate fixes, ownership confusion, or review bottlenecks that threaten a single-mainline delivery model.
The English summary is being prepared. Showing the original description.
Establish build performance baselines and apply systematic optimization techniques. Only activate in MSBuild/.NET build context. USE FOR: diagnosing slow builds, establishing before/after measurements (cold, warm, no-op scenarios), applying optimization strategies like MSBuild Server, static graph builds, artifacts output, and dependency graph trimming. Start here before diving into build-perf-diagnostics, incremental-build, or build-parallelism. DO NOT USE FOR: non-MSBuild build systems, detailed bottleneck analysis (use build-perf-diagnostics after baselining).
The English summary is being prepared. Showing the original description.
Codified expertise for production scheduling, job sequencing, line balancing, changeover optimisation, and bottleneck resolution in discrete and batch manufacturing.
The English summary is being prepared. Showing the original description.