724 skills · sorted by relevance
Check generated PDFs with PyPDF2 before reporting completion. A Python script verifies that each file can be opened and has the expected page count, helping identify outputs that need regeneration.
Define when a file creation task is complete: the deliverable is saved, contains the required content, and uses the correct format and location. Verify those conditions, then report completion without searching for an unnecessary submission tool.
Manage an assigned Worker task from acceptance through delivery using taskflow. Keep work and optional progress notes in the task directory, submit structured results with deliverable paths, and report completion or blockers to the coordinator.
Guides task delegation to Worker agents, completion handling, and recurring task management. It explains how to register assignments, synchronize task files and results, and choose the relevant procedure when checking worker availability or updating task state.
Use this skill when the user is doing hands-on DOCA RDMI (RDMA Initiator) programming — picking doca-rdmi vs doca-rdma for an accelerator-initiated one-sided RDMA flow, standing up a doca_rdmi_connection or doca_rdmi_poster, attaching a doca_dpa_completion or doca_verbs_cq before doca_ctx_start(), retrieving the DPA-side handle for a DPA kernel, auditing whether a doca_rdmi_* symbol is EXPERIMENTAL on this DOCA, or debugging DOCA_ERROR_* returns from RDMI calls. Trigger even when the user does not say "DOCA RDMI" or "initiator" — implicit phrasings include "my DPA kernel needs to post RDMA writes to a remote responder", "DPA kernel sees no completions", "function not found on doca_rdmi_* at link time", "DOCA_ERROR_BAD_STATE from completion attach", or "the DPA posted but the work request never arrived". Refuse and route elsewhere for two-sided or host-CPU RDMA, the DPA programming model, GPU-side RDMA initiation, or general RDMA/IB/RoCE concepts — those belong to other skills.
The English summary is being prepared. Showing the original description.
Expert collection-loop systems for collectible IDs, scavenger hunts, completion archives, nearest-item compass UI, hidden spawners, and persistent find-all-X progress. Use when implementing collectibles, scavenger hunts, completion% archives, or compass-guided item hunts. Keywords: collectible_id, scavenger_hunt, collection_manager, collection_compass, completion_archive, hidden_item_spawner, find_all, collectible_item.
The English summary is being prepared. Showing the original description.
Expert collection-loop systems for collectible IDs, scavenger hunts, completion archives, nearest-item compass UI, hidden spawners, and persistent find-all-X progress. Use when implementing collectibles, scavenger hunts, completion% archives, or compass-guided item hunts. Keywords: collectible_id, scavenger_hunt, collection_manager, collection_compass, completion_archive, hidden_item_spawner, find_all, collectible_item.
The English summary is being prepared. Showing the original description.
Plan, implement, or review verification before completion work in an existing codebase with compatibility, security, and verification controls. Use when the user explicitly requests verification before completion work.
The English summary is being prepared. Showing the original description.
Use when work produces a deliverable, factual claim, dataset, code change, decision, publication, deployment, automation, or irreversible action whose failure would matter; when the user asks for a quality gate, acceptance criteria, QA, completeness, validation, audit, evidence, preflight, release readiness, or a definition of done; or before claiming completion on medium- or high-risk work. Automatically decide whether a formal gate is warranted, derive task-specific pass/fail criteria, gather evidence, and block unsupported completion. Make sure to use this skill even when the user does not say "quality gate" if consequential work needs acceptance criteria or completion evidence. Skip formal gating for trivial, reversible, low-impact requests.
The English summary is being prepared. Showing the original description.
Inspect and control interactive command-line programs running in tmux, a terminal session manager. List sessions and panes, capture current output or history, send text and special keys, and watch for prompts or completion markers.
Checks whether work is truly complete before reporting success, committing changes, or creating a pull request. It runs the relevant checks, reads their full output, and ties status reports to current evidence.
Create and edit documents that AI agents follow, including skills and project instructions. Organize steps and reference material, define clear completion criteria, and refine document links and wording to reduce duplication and unnecessary context.
Turn a draft request into a ready-to-paste prompt by clarifying intent, missing context, and completion criteria. Recommend suitable ECC commands, skills, and agents, then provide full and compact versions with explanations for the changes.
Check learning-log timestamps, free disk space, and shortcut-related phrases before Claude Code ends a session. A Stop hook, which runs when the agent tries to finish, warns or blocks completion according to configured rules without using AI judgment.
Set up and optimize large language model inference, or response generation, on NVIDIA GPUs. Use quantization, batch processing, and multiple GPUs to improve serving performance, and expose models through a chat completion API.
Handle tasks managed by Paperclip by reading their context, saving documents and deliverable files, updating progress and blockers, and coordinating delegated work. The skill provides procedures for short agent runs, approvals, and traceable task updates.
Generate music through Venice.ai, handling queued requests, completion, and result retrieval. Use it to create short jingles, repeating background music, or provisional scores for prototypes as part of an agent-assisted music workflow.
Organize new features, complex implementations, and architecture changes into the SPARC stages: specification, pseudocode, architecture, refinement, and completion. Define requirements and acceptance criteria, consult existing patterns, and record lessons for future work.
Track objectives across multiple conversations by defining milestones, saving progress and findings, and planning each session. Review completion criteria and flag schedule risks, scope changes, or dependencies that may require a different approach.
Send a single request to DeepSeek for summarization, information extraction, or simple classification. Receive the response and token usage as JSON for further processing, with structured status information when the API is unavailable.
Send proofs, planning questions, or difficult debugging tasks to DeepSeek's reasoning model. Receive the final answer and model-provided reasoning text in separate fields, with token usage reported for reviewing responses and costs.
Choose the next task using patterns learned from completed work and current progress. The skill acts on confident predictions, falls back to task priorities when confidence is low, and updates its learning after completion.
Breaks long-running goals into small, verifiable steps, tracks progress, and records evidence from actual operation. Checkpoints preserve what has been proven so work can resume after interruptions and continue toward explicit completion criteria.
Integrate FLUX image generation and editing into applications through the BFL API. The guide covers authentication, checking asynchronous results, completion notifications through webhooks, request limits, error handling, regional endpoints, and downloading generated images before their URLs expire.