263 skills · sorted by relevance
Create shareable links that open specific Arize resources, including traces, datasets, and evaluators. Use IDs from logs or exported data, with suitable time ranges, to direct teammates to the relevant view.
Investigate SGLang serving failures by collecting server state, preserving failing requests, and replaying them on a clean target. Summarize evidence and likely causes, then choose the next debugging tool for crashes, stalls, or performance regressions.
Build Python applications that keep other tasks moving while waiting for network, file, or database operations. The skill guides asyncio patterns, task cancellation, timeouts, workload limits, and testing for asynchronous APIs, scrapers, and background processing.
Coordinates specialist agents in Claude Code to work from product requirements through design, implementation, testing, deployment, and operations. Includes task queues, saved progress, parallel reviews, and recovery workflows for managing extended development work.
Draft prospect emails and follow-up sequences using the sender’s established writing style and specific facts about each recipient. Present messages for approval, prepare approved sends when connected, and record contact history in HubSpot or a file.
Prepare overdue invoice reminders using accounting records and recent payment data. Adjust each message to the customer's payment history, flag possible payments for checking, and send through PayPal or queue email drafts after approval.
Prepare legal work briefings from connected email, calendars, documents, and contracts. Choose a daily overview, an internal topic review, or an incident summary, with urgent items, relevant background, information gaps, and suggested next steps.
Analyze business workflows to find delays, repeated work, unnecessary steps, and opportunities for automation. Get a before-and-after process comparison, specific recommendations, estimated effects, and an implementation plan based on how the work currently happens.
Size support and operations teams using demand ranges and queueing calculations, which estimate how work waits for service. Assess workload imbalance and service target risks, then build a quarterly hiring plan that accounts for training time and attrition.
Use the Archon CLI to launch AI workflows on projects, inspect their results, and approve, reject, cancel, or resume runs. The skill also guides setup, configuration changes, workflow creation, and improvements to workflow prompts.
Manages security incident cases in Google SecOps, from reviewing queues and related alerts to updating priorities, assignments, and investigation notes. It supports closing cases with documented causes, with available actions depending on the connected MCP tools.
Design iii backends by combining functions, queues, and shared state. Choose patterns for workflows with retries, live updates, cooperating AI agents, and event-driven processing, with examples showing how the pieces connect.
Automate tmux, a tool for managing multiple terminal workspaces, to inspect output, send commands, monitor stalled tasks, and record logs. Coordinate multiple AI terminals with task queues, file locks, status tracking, and handoff reports.
Build Laravel applications with authenticated APIs, interactive interfaces, and background jobs. The skill guides database relationships, query optimization, and feature tests, using tools such as Sanctum for authentication and Livewire for interfaces.
Classify agent work as ready to continue, needing human judgment, or intentionally paused. Structure decision requests with evidence, a recommended choice, and clear outcomes, then connect the recorded decision to the next responsible agent.
Execute an existing Ouroboros Seed, a specification containing a goal and acceptance criteria, as a background job. The workflow checks each criterion, normally queues formal evaluation, and follows a bounded improvement loop if evaluation rejects the result.
Find the relevant architectural guidance for API design, security, cloud infrastructure, message queues, and caching. This skill acts as an index to specialized references and lists principles for structuring software, distributed systems, and access controls.
Add OpenTelemetry tracing to frontend and backend flows. Connect execution records across services and background jobs, record phase durations and entity IDs, and mark failures so developers can investigate slow or broken operations.
Deploy teams of AI agents on Flow Nexus and automate development, research, or data processing workflows. Define step dependencies and event triggers, assign suitable agents, and monitor progress, queues, and execution history.
Administer Observal users, settings, submission reviews, identity configuration, and security investigations. Also manage local server upgrades, rollback, and database schema migrations, checking the resulting state and protecting sensitive values throughout privileged operations.
Read WeChat Official Account articles through available web or browser tools. Summarize the content actually accessed, check article details and relevant images, and clearly report reading limits, verification barriers, and claims that remain unverified.
Operate the DOCA UROM Service container on BlueField Arm for remote memory operations (puts, gets, atomics, collectives) enqueued by a paired host using `doca-urom`: pull the NGC image, choose the UCX component, size queues, configure Comch pairing, and align host and service versions. SECURITY: the service has no standalone access control; Comch pairing and RDMA permissions are the boundary. Pair only intended hosts, expose least-privilege memory regions, and verify both views before start. Trigger for slow UCX collectives, unexpected NOT_PERMITTED, or missing completions. Do not use for host application code, MPI/UCX integration design, or DOCA install.
The English summary is being prepared. Showing the original description.
DOCA Rivermax (doca_rmax) receive streams for timing-precise media over IP (SMPTE ST 2110 video / audio, market data, scientific feeds) on a BlueField DPU or ConnectX host, once the Rivermax SDK and its license are installed; getting the SDK or license is doca-setup / doca-public-knowledge-map territory. Trigger on "ST 2110 receive isn't getting frames", "sub-microsecond jitter on my media receiver", "NOT_SUPPORTED from doca_rmax_init", "no recv events after stream start", or "license check failing on a media receiver", even without naming Rivermax. Covers doca_rmax_in_stream sessions, the SDK + license precondition check, doca_rmax_get_*_supported capability queries, pairing with doca-eth queues and doca-flow steering, and DOCA_ERROR_* from Rivermax calls. Refuse and route elsewhere for programming the underlying queue (doca-eth), steering rules (doca-flow), PTP / clock sync (doca-firefly), or best-effort packet I/O.
The English summary is being prepared. Showing the original description.
Use this skill for hands-on DOCA GPI programming — wiring a GPU-Packet-Initiator context so a CUDA kernel drives RDMA queues directly from GPU memory without host CPU mediation. Covers picking GPI vs doca-gpunetio, the doca_gpi / domain / channel object model, the GPU-side handle handoff (doca_gpu_gpi_channel*), attaching GPU memory to a GPI domain, the domain and channel attribute objects, and debugging DOCA_ERROR_* from doca_gpi_* calls. Trigger even when the user does not explicitly mention "DOCA GPI" — implicit phrasings include "my CUDA kernel needs to post RDMA directly from GPU memory", "DOCA_ERROR_* from doca_gpi_gpu_channel_get", "how do I hand a GPU handle to my CUDA kernel", "how many channels can a GPI domain hold", or "GPU kernel driving RDMA without the host CPU on the path". Refuse and route elsewhere for the doca-gpunetio Send/Receive surface, the doca-rdma queue lifecycle, DPA-side initiation (doca-rdmi), or the CUDA programming model — those belong to other skills.
The English summary is being prepared. Showing the original description.