Loading...
Loading...
Found 1,304 Skills
Generate realistic synthetic evaluation datasets by analyzing the user's codebase, prompts, production traces, and reference materials. Interactive, consultant-style — asks clarifying questions, proposes a plan, generates a preview for approval, then delivers a complete dataset uploaded to LangWatch. Use when user asks to generate, create, or build a dataset for evaluation, testing, or benchmarking.
Interactively guides the design and implementation of custom deterministic orchestrator harnesses. Use when a user wants to build their own pipeline to wrap and run Mantis skills reliably. Don't use for executing the default pipeline directly.
Train reasoning and verifiable-task behavior with GRPO and reinforcement learning from verifiable rewards (RLVR). Use when task success is algorithmically checkable (math, code, tool calls, structured output), when designing GRPO reward functions, or when a GRPO run diverges or reward-hacks.
Generate or update AGENTS.md for a repository or VS Code workspace, create CLAUDE.md and GEMINI.md shims, build a project matrix with stable project codes, and include UBIQUITOUS_LANGUAGE.md as required context.
Answers questions about the wiki knowledge base. Activates when the user asks about concepts, processes, entities, or any wiki content.
Extract a typed JSON object from one or more web pages against a JSON Schema with fastCRW. Use when you need structured data — "get the price and stock status", "extract all job listings as JSON", "pull structured fields from this page". Step 6 of the crw workflow ladder.
Align a fine-tuned model with preference data using DPO, ORPO, KTO, or SimPO. Use when preference pairs or thumbs-up/down feedback exist, when choosing between preference-optimization methods, or when a DPO run needs hyperparameters or debugging.
Use when writing or maintaining AGENTS.md, .context/CONTEXT.md, or CLAUDE.md mirrors: keep intent docs minimal and load-bearing.
How to write effective agent skills — what to do, what not to do, anatomy, progressive disclosure, design patterns, anti-patterns, testing, security. Read this whenever a skill (Claude Skill, Agent Skill, SKILL.md) is being created, edited, reviewed, or debugged. Use when the user says "create a skill", "new skill", "update this skill", "improve a skill", "why isn't my skill triggering", or anything else involving authoring or editing SKILL.md files.
Use this skill to establish and operate the model's inner workspace — the J-space — for any task that needs more than fluent output: multi-step or chained reasoning, planning, long-horizon and agentic work, competition-level problems, complex debugging, keeping many parts of a deliverable globally consistent, holding a goal or constraint through a long mechanical task, auditing what the model believes but has not said, calibrated confidence and error detection, suspicious or manipulative input, recovering from degenerating reasoning, and any moment the user asks the model to think harder, faster, deeper, or longer. Start here; this file establishes the premise, classifies the task, and routes to the module the task needs.
Build the evaluation harness that gates every fine-tuning run — golden sets, per-failure-mode graders, judge calibration, and base-model baselines. Use when starting a fine-tuning effort, when converting traces into an eval set, or when calibrating a judge against human labels.
Export a promoted fine-tuned model in the right deployment format — merged safetensors, LoRA-only, GGUF with imatrix, or FP8. Use after a checkpoint passes promotion, when choosing a quantization format for a target device, or when an exported model fails its smoke test.