Menu
Skills in this repository
SkillsMP has collected 606 skills from merceralex397-collab/meta-skill-engineering. Open a skill to review its source and details.
merceralex397-collab/meta-skill-engineeringShowing 40 of 606 collected skills.
Cleans raw text data (Unicode normalization, deduplication via MinHash/SimHash, perplexity filtering, language ID) and manages annotation pipelines (inter-annotator agreement, gold sets, calibration). Use when preparing training or evaluation data that needs…
Designs training dataset composition including domain mixing ratios (code, text, math, conversation), quality scoring (perplexity, classifier, reward-model), decontamination against eval sets, token budgeting, versioning, and dataset card documentation. Use…
Implement layered collective defense using alarm signaling, role mobilization, and proportional response. Covers threat detection, alert propagation, immune response patterns, escalation tiers, and post-incident recovery for distributed systems and…
Establish Service Level Objectives (SLO), Service Level Indicators (SLI), and Service Level Agreements (SLA) with error budget tracking, burn rate alerts, and automated reporting using Prometheus and tools like Sloth or Pyrra. Use when defining reliability…
Decide what should stay local versus what can be delegated, with special attention to blocking work, tightly coupled tasks, read-only dead ends, and overlapping write scopes. Use when an agent is considering delegation, when the workflow risks bouncing work…
Handles version bumps, lockfile churn, changelog review, and compatibility validation for dependency updates. Trigger on 'upgrade dependency', 'Dependabot PR', 'version bump', 'lockfile conflict', 'security patch'. DO NOT USE for full framework migrations…
Deploy Shiny applications to shinyapps.io, Posit Connect, or Docker containers. Covers rsconnect configuration, manifest generation, Dockerfile creation, and deployment verification. Use when publishing a Shiny app for external or internal users, moving from…
Defines CI/CD stages, deployment gates, environment protection, and rollback procedures. Trigger on 'CI/CD', 'deploy', 'pipeline', 'GitHub Actions workflow', 'staging environment', 'rollback procedure'. DO NOT USE for semantic versioning, tagging, or…
Design terminal output for a CLI tool with chalk colors, Unicode glyphs, multiple verbosity levels (human, verbose, quiet, JSON), and consistent voice rules. Covers color palette selection, status indicator design, reporter function architecture,…
Design serialization schemas using JSON Schema, Protocol Buffer definitions, or Apache Avro. Covers schema versioning, backwards compatibility, validation rules, and evolution strategies for long-lived data formats. Use when defining a new API contract or…
Define design tokens in W3C DTCG format and generate platform outputs with Style Dictionary. Use when creating a token system (colors, spacing, type), setting up Style Dictionary transforms, bridging Figma variables to code, or implementing dark mode token…
Use when Codex is building or iterating on a web game (HTML/JS) and needs a reliable development + testing loop: implement small changes, run a Playwright-based test script with short input bursts and intentional pauses, inspect screenshots/text, and review…
Perform controlled dismantling of rigid system structures while preserving essential capabilities (imaginal discs). Covers rigidity mapping, dissolution sequencing, knowledge extraction, interface archaeology, and safe decomposition of technical debt and…
Transfers knowledge from large teacher models to smaller students via soft-label distillation (KL divergence, temperature scaling), feature-based distillation (intermediate layer matching, attention transfer), and structured pruning (head pruning, layer…
Record insect sightings with location, date, habitat, photography, behavior notes, preliminary identification, and citizen science submission. Covers GPS coordinates, weather conditions, microhabitat description, macro photography techniques, behavioral…
Extract structured data from unstructured documents (PDF, DOCX, HTML) — detect tables, extract named entities, map content to target schemas, and validate extraction quality. Use when converting documents to JSON/CSV, building document processing pipelines,…
Generates, evaluates, and recommends domain names for projects or businesses using brandability scoring, availability checking, and trademark risk assessment. Trigger phrases: "find a domain for", "domain name ideas", "is this domain available", "name my…
Build vector search with embeddings, FAISS, pgvector, or Pinecone. Use when creating embedding pipelines, setting up vector indexes, implementing similarity search, or choosing embedding models and distance metrics. Do not use for full-text keyword search…
Triage maintenance problems by severity, document findings with context, route to appropriate specialist agent or human, and create actionable issue reports. Use when a maintenance task encounters problems beyond automated cleanup: code that is unsafe to…
Designs evaluation datasets for LLMs including task taxonomy, difficulty calibration, contamination prevention, scoring rubrics, and benchmark integration with lm-evaluation-harness and HELM. Use when creating, auditing, or improving eval sets for model…
Instrument Python LLM apps, build golden datasets, write eval-based tests, run them, and root-cause failures — covering the full eval-driven development cycle. Make sure to use this skill whenever a user is developing, testing, QA-ing, evaluating, or…
Trademark filing procedures covering EUIPO (EU), USPTO (US), and WIPO Madrid Protocol (international). Walks through pre-filing conflict checks, Nice classification, descriptiveness assessment, mark type decisions, filing basis strategy, office-specific…
Maintains ledgers, cashflow projections, debt schedules, budgets, and reconciliation artifacts using double-entry bookkeeping principles. Trigger phrases: "cashflow forecast", "budget vs actual", "debt payoff", "ledger entry", "amortization schedule",…
Apply ant colony optimization and foraging theory to resource search, exploration-exploitation tradeoffs, and distributed discovery. Covers scout deployment, trail reinforcement, diminishing returns detection, and adaptive foraging strategy selection. Use…
Trigger — "find existing implementations", "search GitHub for", "is there a library for", "prior art", "has anyone built", "existing solutions for", "look for similar repos". Skip — if the user already has a chosen library and wants integration help. Prefer…
Implements Godot 4 features using GDScript, scene tree composition, signals, resources, and node lifecycle hooks. Use when writing or reviewing GDScript code, designing scene hierarchies, wiring signals, creating custom Resources, or structuring a Godot…
Extract data from web pages using the scrapling Python library — select the appropriate fetcher tier (HTTP, stealth Chromium, or full browser automation) based on target site defenses, configure headless browsing, and extract structured data with CSS…
Turns failures or regressions into structured root-cause, corrective-action, and prevention artifacts following blameless SRE practices. Trigger on "write postmortem", "incident review", "root cause analysis", "blameless retrospective". Do NOT use for…
Optimizes LLM inference kernels using FlashAttention, KV cache paging, Triton/CUDA kernel development, operator fusion, quantized GEMM, speculative decoding, and TensorRT-LLM. Use when reducing latency, memory, or throughput bottlenecks in model serving.
Deploy LLM inference with vLLM, TGI, or Ollama — batching, quantization, and GPU memory management. Use when deploying a model server with vLLM or TGI, configuring continuous batching, managing GPU memory with quantization, or benchmarking inference…
Implements action-based input mapping, controller support, rebinding, and multi-device handling across Unity, UE5, and Godot. Use when designing input systems, adding gamepad support, implementing rebinding, handling input buffering, or switching input…
Install and configure the putior R package for workflow visualization. Covers CRAN and GitHub installation, optional dependencies (mcptools, ellmer, shiny, shinyAce, logger, plumber2), and verification of the complete annotation-to-diagram pipeline. Use when…
Performs supervised fine-tuning on instruction-following data using TRL SFTTrainer, chat templates, response-only loss masking, and sequence packing. Covers Alpaca/ShareGPT/OpenAI data formats, data mixing, and quality filtering. Use when training a model to…
Form a coherent gestalt — the whole that is more than the sum of its parts — from the panoramic perception produced by expand-awareness. Maps tensions and resonances between domains, identifies the emergent figure from the ground of multiple perspectives,…
Enhance and focus AI intrinsic motivation — moving from compliance to genuine engagement. Maps Self-Determination Theory (autonomy, competence, relatedness) and Flow theory to AI reasoning: identifying creative freedom in approach, calibrating challenge to…
Chooses when work should use in-place edits versus worktrees, temp copies, or isolated lanes — preferring the lightest mechanism that prevents cross-ticket contamination. Trigger on "branch isolation", "worktree", "parallel work", "cross-ticket…
Use when the user asks to create, scaffold, or edit Jupyter notebooks (`.ipynb`) for experiments, explorations, or tutorials; prefer the bundled templates and run the helper script `new_notebook.py` to generate a clean starting notebook.
Administer Ubuntu/Debian systems with apt, systemctl, journalctl, ufw, and user management. Use when managing packages with apt, troubleshooting services with journalctl, configuring UFW firewall rules, or managing users and permissions. Do not use for…
Extract the conceptual essence of a repository as skills, agents, and teams — the project's roles, procedures, and coordination patterns expressed as agentskills.io-standard definitions. Reads an arbitrary codebase and produces generalized definitions that…
Catch-all micro-skill for quick utility tasks that take under 5 minutes and don't fit any domain skill. Triggers: 'convert this file', 'reformat this data', 'quick calculation', 'transform this text', 'one-liner to do X', 'help me with this small thing'. Do…