Loading...
Loading...
Found 4,751 Skills
How to launch distributed Megatron-LM training jobs on a SLURM cluster. Covers a minimal sbatch skeleton, environment-variable setup for torch.distributed.run, CUDA_DEVICE_MAX_CONNECTIONS rules across hardware and parallelism modes, container conventions, monitoring, and per-rank failure diagnosis.
Set up systematic review monitoring across e-commerce platforms. Track new reviews, detect negative review spikes, monitor competitor reviews, and automate review response workflows.
Use when the user has a long-form video (interview / lecture / podcast / conversation) and a transcript SRT, and wants to extract 3–6 stand-alone topical short clips from it. This skill ONLY cuts and crops — it produces raw clips + per-clip SRTs as a hand-off package for downstream post-production (`/wjs-overlaying-video`). Triggers — "切成几段", "分主题", "拆成短视频", "切片", "topic segments", "split into clips".
Multi-model deep review of the Ralph bd graph and plan via three parallel opencode processes (claude opus, gemini, gpt). Use for high-stakes runs where cross-model consensus reduces single-model bias.
Run large codebase migrations and multi-file refactors. Uses the Composio CLI to coordinate issue tracking, batched PRs, and CI verification while the agent executes the transforms locally across hundreds of files.
Query and filter Datadog logs from the shell using the Composio CLI. Run scoped log searches, pivot across services/environments, and export structured JSON for downstream agents instead of click-driving the Datadog UI.
Tests WebSocket API implementations for security vulnerabilities including missing authentication on WebSocket upgrade, Cross-Site WebSocket Hijacking (CSWSH), injection attacks through WebSocket messages, insufficient input validation, denial-of-service via message flooding, and information leakage through WebSocket frames. The tester intercepts WebSocket handshakes and messages using Burp Suite, crafts malicious payloads, and tests for authorization bypass on WebSocket channels. Activates for requests involving WebSocket security testing, WS penetration testing, CSWSH attack, or real-time API security assessment.
Use this skill when the user says "/li", or when the request is related to content creation but the intent is ambiguous and the user is unsure which tool to use. As the main entry point of the li Toolkit, it judges the intent and routes to the corresponding specialized skill. Do NOT trigger: When the user's request clearly matches a specific li-* skill (e.g., "write a script" → li-writer, "deepen a topic" → li-topic), directly trigger the corresponding skill instead of this entry. DO NOT trigger for non-content-creation tasks. Use when the user says "/li" or the intent is ambiguous across multiple li-* tools.
QLC+ (Q Light Controller Plus) lighting software — workspace files (.qxw), scenes, chasers, sequences, collections, EFX, RGB matrices, fixture definitions (.qxf), Virtual Console, and timing calculations. Use this skill whenever the user is working with QLC+ workspace XML, creating/editing scenes/chasers/shows, debugging timing or crossfade issues, generating fixture definitions, setting up Virtual Console widgets (cue lists, solo frames, buttons, sliders), troubleshooting HTP/LTP conflicts, fixing corrupted .qxw files, configuring QLC+ plugins (DMX USB, Art-Net, MIDI, E1.31), or asking any question where QLC+ is the software being used. Also trigger for SpeedModes, FadeIn/Hold/Duration, FixtureVal, RunOrder, or QLC+ function types. Do NOT trigger for general DMX hardware questions without QLC+ context, fixture buying advice, DAW-only questions, or custom protocol implementations.
Build and maintain a Karpathy-style LLM knowledge base — a self-compiling Obsidian markdown wiki where an Agent ingests raw sources, compiles cross-linked concept/entity/summary pages, answers queries against the corpus, lints the graph for health, and audits in-context human feedback filed from Obsidian or the local web viewer. Use when (1) scaffolding a new knowledge base for any research topic, (2) ingesting articles/papers/PDFs/web pages into raw/, (3) compiling or restructuring wiki articles from existing raw material, (4) answering questions against the wiki and filing durable answers back, (5) running lint passes for dead links / orphan pages / coverage gaps / audit shape, (6) processing human feedback from the audit/ directory and applying corrections. Not for general note-taking, daily journals, or non-wiki Obsidian use.
Build and maintain a research repository that makes findings findable, reusable, and cumulative across the organization.
Triage a subpoena served on the company — classify it, analyze scope/burden/privilege, cross-check the portfolio, and produce an objections framework, compliance plan, and deadline calendar. Use when the user says "we got a subpoena", "served with a subpoena", or shares a subpoena, CID, or third-party document request to evaluate.