Skills
97,501 packaged capabilities and playbooks for agents. Install any of them with one command.
Data & databases
Use when reviewing PL/SQL - a changed procedure or package, a proposed file, or an object suspected of causing trouble. Combines the database's own correctness signals with the antipattern checklist, and reports findings anchored to line numbers.
Use when writing or modifying PL/SQL source - a procedure, function, package, trigger, or view - after impact is known and the spec is settled (open decisions go through pythia-spec first). The codebase's conventions already exist; copy them instead of inventing style, and anchor every type to the database's reality.
Use when answering questions about a dbt v2 project's own metadata — which models, sources, tests, columns, configs, tags, packages or lineage exist, what is untested or undocumented, what depends on what, how long models took in the last run — or when using `dbt show --info`, `{{ info_schema() }}`, `--generate-info-schema`, `target/info_schema/`, or writing `dbt check` SQL. Prefer this over grepping YAML or parsing `manifest.json` on dbt v2.
Read, write, and update Google Sheets data via CLI. Use when the user asks to read spreadsheet data, update cells, append rows, or work with Google Sheets. Triggers on mentions of spreadsheets, sheets, Google Sheets, tabular data in the cloud, or specific sheet names like "Projects" or "Tasks".
Postgres best practices maintained by Supabase, for Postgres running anywhere. Load this skill BEFORE writing or changing anything that lives in a Postgres database: creating or altering tables and columns (including choosing column types), schema design, migrations and declarative schema files, RLS policies and the tests that verify them, indexes, triggers, database functions, queues and scheduled jobs (pg_cron, pgmq), vector/semantic search (pgvector), and restoring dumps (pg_restore) or importing data. Also load it when diagnosing slow queries, high CPU, timeouts, EXPLAIN plans, connection exhaustion, locking, bloat, or rows visible to the wrong user or tenant. This is not just a performance guide — schema, migration, security, and SQL authoring tasks need these rules too, even for a one-column change or a single query.
WeWrite data review module: pulls read/share/like data for WeChat Official Account articles, backfills historical records, and suggests adjustments to topics, titles, and outlines. Trigger keywords: 看看文章数据, 文章数据怎么样, 效果复盘, 看看表现, 阅读量怎么样. Requires WeChat Official Account/WeWrite context; should not be triggered by generic "data analysis".
Guide for writing performant ClickHouse queries in PostHog product code. Use when writing HogQL query runners, designing a ClickHouse table for a new product, adding materialized columns or skip indexes, or choosing a row ID format. For optimizing an existing query that is already too slow, use `/optimizing-clickhouse-and-hogql-queries` instead.
Collect validated Xiaohongshu image assets from normalized XHS datasets into local manifests and downloaded files. Use this when you need reproducible local media artifacts from note covers or other already-exposed remote asset URLs.
Generates structured experimental designs (factorial, response surface, Taguchi) to systematically discover how multiple factors affect outcomes while minimizing experimental runs. Use when optimizing multi-factor systems with limited experimental budget, screening many variables to find the vital few, discovering interactions between parameters, mapping response surfaces for peak performance, validating robustness to noise factors, or when users mention factorial designs, A/B/n testing, parameter tuning, or process optimization.
Calculates probability-weighted averages of all possible outcomes to enable rational decisions under uncertainty. Covers scenario identification, probability estimation, payoff quantification, and risk-adjusted interpretation. Use when comparing risky options (investments, product bets, strategic choices), prioritizing projects by expected return, assessing whether to take a gamble, or when user mentions expected value, EV calculation, risk-adjusted return, probability-weighted outcomes, or decision tree.
Evaluate the financial health of an A-share company through earnings quality, cash conversion, working capital, leverage, and balance-sheet risk. Use this skill when revenue and profit trends diverge from cash flow, when debt and capex matter, or when impairment and receivable quality are major concerns. All outputs must include accessible source links.
Set up and validate Splunk HTTP Event Collector (HEC), explain indexer acknowledgment and distributed HEC behavior, diagnose HEC no-data and HTTP delivery failures from sanitized evidence, and prepare bounded escalation handoffs. Use for Splunk Cloud Platform or Splunk Enterprise HEC tokens, endpoints, event or raw payloads, TLS, channels, ACK, health, authorization, queues, and delivery verification; do not use for non-HEC ingestion, broad architecture, allowlist changes, or service-side remediation.
Owns the iteration loop on top of an ML workspace: the `journal/JOURNAL.md` index and the per-experiment `journal/NN_short_name.md` design notes that must be drafted and approved by the user **before** `experiments/NN_short_name.py` is created. Drives the propose → iterate → approve → implement → record loop; dispatches to `iterate-from-skore` / `iterate-from-user` for sourcing. TRIGGER — any of: - A session opens in an ML workspace (whether or not `journal/` exists yet — missing/placeholder → bootstrap mode). - User says "what's next", "resume", "where were we", "let's iterate", "propose next", "first baseline". - About to create a new `experiments/NN_*.py` (the matching `journal/NN_*.md` must exist and be approved first). - User wants to record an outcome from a finished run. - User asks to compare past experiments or review what's been tried ("compare X and Y", "where are we?"). SKIP when: no `journal/` yet AND no workspace scaffold (route to `organize-ml-workspace`); the work is mechanical inside `pipeline.py` / `evaluate.py` / `data.py` with no journal-level implication (owned by `build-ml-pipeline` / `evaluate-ml-pipeline`); the user asks for a symbol lookup (`python-api`); the user is diagnosing a single skore report without a "what next" framing (`evaluate-ml-pipeline`). HOW TO USE: read `journal/JOURNAL.md` first, classify the turn via the **Mode picker** (table near the top), then read only the matching section. Sibling skills open *just-in-time* when a step requires them — do not pre-read all sibling skills at session start. Design notes are the only artifact this skill writes; read, compare, and overview modes don't write.
Generates Python code that evaluates SageMaker models. Supports two evaluation types: LLM-as-Judge and Custom Scorer. Use when the user says "evaluate my model", "run a benchmark", "test model performance", "how did my model perform", "compare models", or similar requests.
Use before proposing or writing any change to an Oracle object - a procedure, package, function, trigger, view, or table. Impact analysis comes first, because what depends on an object decides how careful the change must be, and Oracle invalidates dependents in cascade.
Owns one experiment's smoke test: a small pytest that fits on part of real `data/` and predicts on a disjoint slice with no pre-history buffer. Prediction count must equal predict-grid rows. TRIGGER when an approved experiment needs its smoke test, `pytest tests/smoke/` fails on row count, the user asks why the smoke test is failing, a pipeline edit needs proof, or the experiment script changes pipeline shape. STOP when `status.setup.pending` is non-empty (load `setup-ml-project`), or when there is no approved design or no matching experiment script: explain and send the user to setup or stay in `build-ml-pipeline`. A declined git is not asked again. "Why is smoke failing?" stays here. This action does not cover regression tests or CV interpretation. Do not write `skore.evaluate`. HOW TO USE: `status`, read the journal and experiment, write `tests/smoke/test_NN_*.py`, then `python -m skore_skills smoke run --stem <stem>`. Red JSON means fix the pipeline in `build-ml-pipeline`. Do not loosen the assertion.
Validate a variant-effect predictor (AlphaMissense, ESM-C SAE, ESM logits, EVE, conservation scores, or any per-variant numeric score) against experimental deep mutational scanning (DMS) data. Computes per-variant predictor scores, splits variants into neutral and disruptive groups by DMS effect, runs a Mann-Whitney U test on the predictor scores, and sweeps the stratification thresholds for robustness. Use when you need to know whether a predictor's scores track real functional disruption on a specific protein.
Feishu online spreadsheet (Sheets) operations, including creating, reading, writing, appending data, and managing worksheets. Use when the user mentions Feishu Sheets, online spreadsheets, or electronic spreadsheets (not Bitable/multidimensional tables). Supports creating spreadsheets, writing and reading cell values, appending rows, inserting or deleting rows and columns, and managing worksheets.
Execute read-only T-SQL queries against Fabric Data Warehouse, Lakehouse SQL Endpoints, and Mirrored Databases via the MCP `fabric-sqlendpoint-execute_query` tool. Default skill for any lakehouse data query (row counts, SELECT, filtering, aggregation) unless the user explicitly requests PySpark or Spark DataFrames. Use when the user wants to: (1) query warehouse/lakehouse data, (2) count rows or explore lakehouse tables, (3) discover schemas/columns, (4) generate T-SQL scripts, (5) monitor SQL performance, (6) export results to CSV/JSON. For a Fabric notebook cell (%%sql or other notebook magics), use `spark-authoring-cli`, not this skill. Triggers: "warehouse", "SQL query", "T-SQL", "query warehouse", "show warehouse tables", "show lakehouse tables", "query lakehouse", "lakehouse table", "how many rows", "count rows", "SQL endpoint", "describe warehouse schema", "generate T-SQL script", "warehouse performance", "export SQL data", "connect to warehouse", "lakehouse data", "explore lakehouse".
Answer questions using the Tenzir documentation. Use whenever the user asks about TQL syntax, pipeline operators, functions, data parsing or transformation, normalization, OCSF mapping, enrichment, lookup tables, contexts, packages, nodes, platform setup, deployment, configuration, integrations with tools like Splunk, Kafka, S3, Elasticsearch, or any other Tenzir feature. Also use when the user asks how to collect, route, filter, aggregate, or export security data with Tenzir, or needs help writing or debugging TQL pipelines, even if they don't mention 'Tenzir' explicitly but are clearly working in a Tenzir context.
Propose the mechanism by which a missense variant causes loss of function (LoF), synthesizing evidence from 5 independent layers: AlphaMissense pathogenicity, AlphaFold structural context, ESMC sequence likelihood, SAE feature disruption, and DynaMut2 stability ΔΔG. Distinguishes “structural stability LoF” (misfolding) from “direct functional disruption” (catalytic / binding / PTM site damage). Use for coding missense variants when you need a mechanistic causal model, not just a pathogenicity score.
Advanced Twitter search and social media data analysis. Fetches tweets by keywords using Twitter API, processes up to 1000 results, and generates professional data analysis reports with insights and actionable recommendations. Use when the user requests Twitter/X social media search, social media trend analysis, tweet data mining, social listening, influencer identification, topic sentiment analysis from tweets, or any task involving gathering and analyzing Twitter data for insights.
Stand up a self-hostable, typo-tolerant search environment with Typesense — the open-source Algolia / ElasticSearch alternative (single C++ binary, <50ms instant search, no runtime deps). One routing-first skill: pick a server mode (binary download, official Docker image, or managed Typesense Cloud), install an API client (Python/JS/PHP/Ruby official; Go/Dart/C# community), design a collection schema, index documents, and run searches with typo tolerance, faceting/filtering, geo-search, sorting, grouping, synonyms, curation, scoped API keys, and federated multi-search — then wire an InstantSearch.js UI and a Raft-based HA cluster for production. Use when the user wants to build or operate an installable search backend, add site/app/product search, or migrate off Algolia/Elasticsearch. Triggers on: typesense, search engine, typo-tolerant search, algolia alternative, elasticsearch alternative, instantsearch, faceted search, geo search, vector search, self-hosted search, site search, product search.
WeChat Official Account article aggregator. Uses the mptext.top API to batch-fetch the latest article lists from specified accounts, download article content, and parse it into Markdown/HTML/plain text. Supports fetching by account name or fakeid and includes 8 popular AI technology accounts. Use cases: retrieving the latest articles from specified accounts, batch scraping content from multiple authors, aggregated reading of account articles, and collecting tech news. Keywords: official account article retrieval, WeChat article scraping, account aggregation, article collection, mptext API.
Use when an agent must act as a World Cup predictor for World Cup predictions, 2026 World Cup simulations, World Cup win/draw/loss probabilities, 90-minute match probabilities, explanations of champion or qualification probabilities, audited offline input checks, continuing a tournament based on completed results, or cautious China football lottery 3/1/0 reference lists
Use when writing or reading GenVarLoader (gvl) datasets — preparing VCF/PGEN/SVAR variant sources with bcftools/plink2, calling gvl.write, configuring gvl.Dataset for haplotype/reference/annotated/variants output modes, attaching BigWig or Table tracks, setting up spliced haplotypes from a GTF, choosing track insertion-fill strategies for indels, or filtering variants by allele frequency.
Generates a 3-month SEO performance HTML report for any domain using DataForSEO data. Fetches baseline vs current traffic, keyword rankings, top content pages, and competitive landscape, then outputs a report styled with the Infrasity brand design system: #0D0A1A background, #8157F2 purple accent, Instrument Sans + DM Sans + DM Mono typography. Use this skill whenever a user provides a target domain and a list of competitor URLs and asks for an SEO report, performance report, SEO analysis, competitive SEO comparison, traffic report, ranking report, or 3-month SEO summary. Also trigger when the user says "generate SEO report for X vs Y and Z", "create a performance report", "compare my SEO against competitors", or pastes a domain and asks how it's performing versus the market. Always use this skill for SEO report generation — do not attempt to build the report without following this structured data-fetch and HTML generation workflow.
Set up an autonomous research loop where an agent iterates on experiments, hypotheses, and code on git branches. Use this skill when the user wants to automate ML experiments, set up a research agent that self-improves, needs to run many variations of an experiment without manual intervention, or says "automate experiments", "agent research loop", "self-improving agent", "run hyperparameter search", "autoresearch". Based on Karpathy 28k-like AutoResearch post.
Looker Studio (formerly Google Data Studio) expertise. Build dashboards, design data visualizations, connect data sources, and create marketing reports. Use when the user asks about Looker Studio, Data Studio, marketing dashboards, data visualization, report building, or connecting analytics data sources.
Analyze data using MATLAB. Use when the task involves tables, timetables, time-series data, numeric arrays, sensor matrices, or gridded data — including but not limited to exploring, row filtering, sorting, cleaning, transforming, aggregating, smoothing, padding, trimming, and answering questions about data. MATLAB provides extensive, easy-to-use built-in functions for these workflows with no additional products required.
Use when managing Oracle Autonomous Database on OCI, troubleshooting performance issues, optimizing costs, or implementing HA/DR. Covers ADB-specific gotchas, cost traps, SQL_ID debugging workflows, auto-scaling behavior, and version differences (19c/21c/23ai/26ai).
Create, edit, and convert Excel workbooks (.xlsx/.xls) using Syncfusion XlsIO. Supports two modes — generate C# code for the user's project, or execute a temporary CSX script. Use when the user mentions Excel, xlsx, workbook, template markers, Syncfusion XlsIO, PDF conversion, or custom fonts for Excel-to-PDF or Excel-to-Image conversion.
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.