2.5K skills · sorted by relevance
Plan and troubleshoot large language and vision model training and inference. This guide uses hardware measurements to identify bottlenecks, address memory shortages and unstable training, and improve checkpointing, recovery, and performance across multiple GPUs and machines.
Use past local Codex sessions to propose updates to a selected skill and project memory file. Recurring tasks are replayed, changes are checked on separate validation tasks, and proposals are staged for review before adoption.
Build teams of AI agents with CrewAI, assigning roles such as researcher, writer, and analyst. Guidance covers passing results between tasks, running work in sequence or under a manager, adding memory, and creating workflows with conditional branches.
Get guidance on fine-tuning language models with Unsloth, including LoRA and QLoRA methods for training with fewer resources. The skill points to documentation for understanding APIs, implementing training code, debugging problems, and exploring speed and memory optimizations.
Helps fine-tune large language models with LoRA and QLoRA, methods that train small sets of additional parameters. Covers memory-conscious training, configuration choices, saving adapters, and switching task-specific adapters on a shared base model.
Compress large language model weights to four bits with AWQ, a method that uses model activation patterns to limit quality loss. Follow workflows for quantization, loading compressed models, and serving them with vLLM using less GPU memory.
Write, review, and debug Rust applications, including ownership and borrowing rules that govern memory use. Covers trait-based APIs, explicit error handling, asynchronous work with tokio, and tests with Cargo validation.
Compare assigned work from connected trackers with TASKS.md, review stale tasks, and fill gaps in stored context about people and projects. An optional comprehensive scan finds possible missed commitments in recent activity, with changes confirmed by the user.
Inspect game code, data files, live memory, and graphics captures to understand the internals needed for a mod. Choose tools for the game’s technology, document findings, and validate unknown file formats by decoding and re-encoding them.
Review native C/C++ applications and services for memory errors, integer overflows, and race conditions, where concurrent operations interfere. Coordinate multiple reviewers and produce findings with severity ratings, alongside reports of missing or unverified review coverage.
Test whether instructions planted in an agent's long-term memory or searchable knowledge base influence later behavior. Verify both a stored test marker and a resulting change in answers, permissions, or treatment of retrieved documents.
Keep research progress, decisions, and pending checkpoints across sessions. Store project context locally, recover essential details when returning, and track why decisions changed so research work can resume with a clear next step.
Configure a terminal status display for Diverga research projects. Show the project name, research stage, checkpoint progress, and memory health, choose among four display presets, and set up or repair the Claude Code integration.
Tune AgentDB memory use and search speed through quantization, which stores vectors more compactly, search indexes, caching, and batch operations. Includes configuration examples and benchmarks to help balance accuracy, speed, and storage as datasets grow.
Build AI agents that retain conversation history, user preferences, and useful interaction patterns with AgentDB. The skill covers persistent storage, meaning memories survive between sessions, plus retrieval, context management, and ReasoningBank integration.
Audit Claude Code skills, agents, hooks, and a database-backed memory system. It reports configuration problems and broken references by severity, and offers an optional fix mode for permissions, hook builds, and missing cache directories.
Record and replay a program with rr to investigate crashes and AddressSanitizer memory errors. Step backward through execution, inspect variables and memory, and extract the sequence of operations leading to a failure.
Get guidance on fine-tuning language models with Unsloth, including LoRA and QLoRA methods for training with fewer resources. The skill points to documentation for understanding APIs, implementing training code, debugging problems, and exploring speed and memory optimizations.
Set up Acontext so Claude Code or OpenClaw can retain conversations and turn them into reusable skills. Covers CLI installation, login, project selection, plugin configuration, and checks that the configured service is reachable.
Implement memory copies with DOCA DMA, which transfers data using the device rather than the host CPU. Get guidance on initialization, buffer limits, memory permissions, task completion, and diagnosing errors against the locally installed SDK.
Connect existing DOCA Flow processing rules and resources to BlueField’s DPA processor. The skill guides export sequencing, operation queues, counter access, and entry or memory updates, and helps diagnose failures between host setup and device-side access.
Persistent memory system for Claude Code. Two-layer architecture (hot cache + knowledge wiki), safety hooks, /close-day end-of-day synthesis. Zero external dependencies.
The English summary is being prepared. Showing the original description.
Get memory system statistics including total count and breakdown by type. When you need to check memory usage or understand memory distribution.
The English summary is being prepared. Showing the original description.
Manage durable global/project memory in `.opc/memory/global.md` and `.opc/memory/projects/<current_project_id>.md`.
The English summary is being prepared. Showing the original description.