1.4K skills · sorted by relevance
Configure large language model pretraining with torchtitan across multiple GPUs. Covers ways to split training work, launch jobs on computing clusters, enable Float8 numerical processing, and save or resume training through distributed checkpoints.
Turn feature requirements into a PRD, a document defining what to build, then organize development plans and GitHub Issues. Coordinate parallel agents and track progress, blockers, and next priorities through delivery.
Investigate file access problems by checking folders and permissions, then reviewing files in parallel groups. Audit TypeScript imports, including type-only imports and duplicate statements, and check index exports before summarizing fixes and unresolved issues.
Diagnose file access problems by checking directories, permissions, file headers, and line counts in parallel batches. Also reviews TypeScript imports, fixes duplicates and type-only imports, and checks module exports before retrying the failed operation.
[omh] Accepted plan awaiting implementation: split it into disjoint parallel lanes with per-lane acceptance criteria, verification commands, and owners; prevents two lanes editing the same file. Aliases: ulw. Use when the user says: ultrawork, parallel work, parallel implementation, parallel then integrate, high throughput, coding team, coordinated workers, finish until done.
The English summary is being prepared. Showing the original description.
Compile current scholarly evidence for a scientific manuscript or research brief. Use when the user explicitly asks to gather literature, references, background evidence, competing findings, or a manuscript research packet. Uses Parallel Search by default, Parallel Extract for source verification, Parallel Research for explicitly deep/exhaustive work, optional explicit Parallel Chat, and optional Perplexity only when requested or allowed as a failure fallback.
The English summary is being prepared. Showing the original description.
Help users write correct R code for async, parallel, and distributed computing using mirai. Use when users need to run R code asynchronously or in parallel, write mirai code with correct dependency passing, set up parallel workers, convert from future or parallel, use mirai_map, integrate with Shiny or promises, or configure cluster/HPC computing.
The English summary is being prepared. Showing the original description.
Parallel/distributed computing for larger-than-RAM data. Components: DataFrames (parallel pandas), Arrays (parallel NumPy), Bags, Futures, Schedulers. Scales laptop to HPC cluster. For single-machine speed use polars; for out-of-core without cluster use vaex.
The English summary is being prepared. Showing the original description.
Fast PARALLEL wiki lookup engine over wiki/techniques + wiki/payloads + wiki/tools + wiki/cheatsheets for a surface/service/vuln-class. Two modes - quick (one qmd search, cheap, fire constantly) and deep (4 parallel subagents, one per area, merged ready-to-use arsenal card, cached). This is the fast path arsenal that `arsenal` delegates to; the hunt-* skills each inline their own qmd_query and can hand off here for a parallel lookup. Use for "what do I use against <surface>", "arsenal for <X>", "deep/full arsenal", "tool + payload + technique + cheatsheet for <X>", "fast wiki lookup", "parallel wiki search", any "how do I attack/exploit <service|vuln-class>" where you want the documented tooling + payloads before hand-rolling.
The English summary is being prepared. Showing the original description.
Process multiple documents in bulk with parallel execution. Use when a user asks to batch process files, convert many documents at once, run parallel file operations, bulk rename, bulk transform, or process a directory of files concurrently. Covers parallel execution, error handling, and progress tracking.
The English summary is being prepared. Showing the original description.
Analyzes large tasks for independent subtasks that can be safely parallelized. Produces a DAG-based dispatch plan with dependency ordering and maximum parallelism. Use when a task has 5+ subtasks touching independent files, or the user asks to speed up or parallelize the work.
The English summary is being prepared. Showing the original description.
Orchestrate complex development with AUTOMATIC parallel subagent execution, continuous dispatch scheduling, dependency analysis, file conflict detection, and IOSM quality gates. Analyzes task dependencies, builds critical path, launches parallel background workers with lock management, monitors progress, auto-spawns from discoveries. Use for multi-file features, parallel implementation streams, automated task decomposition, brownfield refactoring, or when user mentions "parallel agents", "orchestrate", "swarm", "continuous dispatch", "automatic scheduling", "PRD", "quality gates", "decompose work", "Mixed/brownfield".
The English summary is being prepared. Showing the original description.
Executes DAG waves with controlled parallelism using the Task tool. Manages concurrent agent spawning, resource limits, and execution coordination. Activate on 'execute dag', 'parallel execution', 'concurrent tasks', 'run workflow', 'spawn agents'. NOT for scheduling (use dag-task-scheduler) or building DAGs (use dag-graph-builder).
The English summary is being prepared. Showing the original description.
Set up parallel agent team for COMPLEX tasks. Creates worktree structure, generates ownership map, and initializes task queue. Use when orchestrator decomposes a task into parallel subtasks, or when asked to "set up team", "parallel setup", or "configure worktrees".
The English summary is being prepared. Showing the original description.
Coordinates staged or parallel jobs across live-verified coding-agent runtimes (Claude Code, Codex, Cursor, and other installed CLIs) and in-session toolkit subagents. Routes each job by capability and risk tier, isolates parallel writers in git worktrees, captures redacted output, resumes interrupted runs, and blocks completion until an independent arbiter verifies the results. Use when work should be split across multiple runtimes or subagents, run as staged or parallel jobs, and reviewed before handoff. NOT for a single-agent task or the in-session 7-phase flow (use mk:cook / mk:workflow-orchestrator).
The English summary is being prepared. Showing the original description.
Guide for optimizing MSBuild build parallelism and multi-project scheduling. Only activate in MSBuild/.NET build context. USE FOR: builds not utilizing all CPU cores, speeding up multi-project solutions, evaluating graph build mode (/graph), build time not improving with -m flag, understanding project dependency topology. Note: /maxcpucount default is 1 (sequential) — always use -m for parallel builds. Covers /maxcpucount, graph build for better scheduling and isolation, BuildInParallel on MSBuild task, reducing unnecessary ProjectReferences, solution filters (.slnf) for building subsets. DO NOT USE FOR: single-project builds, incremental build issues (use incremental-build), compilation slowness within a project (use build-perf-diagnostics), non-MSBuild build systems. INVOKES: binlog MCP server tools (expensive_projects, expensive_targets, project_target_times); falls back to dotnet build -m, dotnet build /graph, binlog replay + grep.
The English summary is being prepared. Showing the original description.
Route user requests to the right skill, decompose complex work into parallel subagents, and manage project phase transitions. Load when the user asks "what should I do next", "which skill should I use", "orchestrate this", "run the full workflow", "split this into parallel tasks", or when a complex request spans multiple skills. Also triggers on "coordinate agents", "parallel execution", "task decomposition", "agent workflow", "what phase am I in", or when the user gives a broad instruction that requires multiple skills in sequence. This is the project's brain — it decides what runs, when, and whether to parallelise.
The English summary is being prepared. Showing the original description.
Executes DAG waves with controlled parallelism using the Task tool. Manages concurrent agent spawning, resource limits, and execution coordination. Activate on 'execute dag', 'parallel execution', 'concurrent tasks', 'run workflow', 'spawn agents'. NOT for scheduling (use dag-task-scheduler) or building DAGs (use dag-graph-builder).
The English summary is being prepared. Showing the original description.
Provides PyTorch-native distributed LLM pretraining using torchtitan with 4D parallelism (FSDP2, TP, PP, CP). Use when pretraining Llama 3.1, DeepSeek V3, or custom models at scale from 8 to 512+ GPUs with Float8, torch.compile, and distributed checkpointing.
The English summary is being prepared. Showing the original description.
Investigate queued or slow OpenClaw CI jobs, the automated checks run during development. Use measured evidence to adjust runner capacity, registration budgets, parallel job limits, test grouping, and placement between GitHub-hosted runners and Blacksmith.
Assign independent bugs or test failures to separate agents that work at the same time. Define each agent’s scope and expected report, then check for conflicting changes and run the full test suite before integrating fixes.
Browse available specialist agents by domain, select up to five, and run them independently on the same task in parallel. Receive a combined report showing each agent’s findings, shared conclusions, conflicting recommendations, and suggested next steps.
Coordinate parallel frontend, backend, or service development around one shared, machine-checkable interface specification. Generate consumer types, validate mocks and real responses against the same specification, and review changes together to prevent mismatched fields and incompatible integrations.
Coordinate large research or generation tasks across a model team. An orchestrator plans and checks the work, workers execute subtasks in parallel, and an advisor reviews the plan and draft before the results are combined.