Claude Skills
The open format is called Agent Skills and works in Claude Code, Codex, Cursor and other agents — most people know it as Claude Skills.
Every Agent Skill we could find on GitHub, deduplicated by content. 79 437 files from 1 744 authors, of which 61 785 are unique — the rest is the same skill repackaged into someone else's repository. For each one: what it weighs in tokens, whether it ships runnable scripts, and which MCP servers it needs.
Maintain Wassette's Keep a Changelog-style CHANGELOG.md by reviewing a focused change, deciding whether it affects users, and updating the [Unreleased] section with concise, categorized, non-duplicate entries.
When a symptom has several plausible causes, rank falsifiable hypotheses and run the cheapest discriminating observation first; prefer least-assumptive survivors only after evidence fit.
When a symptom has several plausible causes, rank falsifiable hypotheses and run the cheapest discriminating observation first; prefer least-assumptive survivors only after evidence fit.
When a symptom has several plausible causes, rank falsifiable hypotheses and run the cheapest discriminating observation first.
When a symptom has several plausible causes, rank falsifiable hypotheses and run the cheapest discriminating observation first; prefer least-assumptive survivors only after evidence fit.
When a symptom has several plausible causes, rank falsifiable hypotheses and run the cheapest discriminating observation first; prefer least-assumptive survivors only after evidence fit.
When a symptom has several plausible causes, rank falsifiable hypotheses and run the cheapest discriminating observation first; prefer least-assumptive survivors only after evidence fit.
When a symptom has several plausible causes, rank falsifiable hypotheses and run the cheapest discriminating observation first; prefer least-assumptive survivors only after evidence fit.
When a symptom has several plausible causes, rank falsifiable hypotheses and run the cheapest discriminating observation first; prefer least-assumptive survivors only after evidence fit.
Use when search or investigation could run forever. Set an explicit good-enough threshold first, then stop at the first option that clears it.
Use when a specific claim may lack grounding. Check evidence boundary, size wrongness cost, then answer, fetch, or abstain — never confabulate.
When a fault is localized and the proximate cause is known but the systemic root is not, chain evidence-linked whys with a counterfactual stop and a countermeasure.
When the right response mode is unclear, classify the cause-effect domain first; decompose disorder.
When a constraint is treated as fixed, separate physics from convention, keep only independently supported primitives, and rebuild the simplest solution that satisfies real constraints.
Deciding what to build or why adoption fails. Recover the progress users hire a solution for under a circumstance, then rank by outcome and competing workarounds.
Use when a selective defect needs IS/IS-NOT difference analysis or a consequential option choice needs must/want weighting and adverse-consequence comparison.
Under genuine uncertainty with no reliable forecast, inventory means, cap downside at affordable loss, act for commitments, and let goals emerge from controllable moves.
Use when longevity of a non-perishable option matters. Treat survival duration as a remaining-life prior, then check domain drift before favoring the proven.
When a claim, doc, test, metric, or assumption conflicts with observed behavior, stop theorizing from the map and verify the live code or data; let territory overrule.
When one mental model leaves a material blind spot on a multi-domain or high-stakes problem, sequence complementary models with named roles and a conflict rule.
Use under time pressure when the situation is still changing and you must act before certainty — cycle Observe→Orient→Decide→Act on ~70% confidence, then re-observe.
When unsure which thinking skill fits, map domain and problem type, then return NONE or one primary skill by default (at most three complementary).
Before committing scarce time, people, or money, name the best forgone use of those resources and the value delta of the chosen path versus that alternative.
When provisioning, setting a limit, or committing an estimate under uncertainty, size a buffer to residual error and the cost of breach—not to the optimistic edge.
Before committing to a plan or launch, assume it already failed and reason backward through concrete causes — convert failure paths into mitigations, gates, and stop checks.
Use when forecasting, estimating, or sizing risk — anchor on base rates, give ranges, update prior→likelihood→posterior on evidence, and factor unmeasured quantities into order-of-magnitude bounds.
For authorized security review of code, auth, or APIs you control, model the attacker, map the attack surface, and report only findings with a reproducible exploit path and verified mitigation.
Before heavy deliberation, classify the decision as cheap or costly to undo; decide two-way doors fast and stage one-way doors to preserve options.
When a change has effects past the immediate fix—incentives, scale, feedback—trace consequence chains with timing and probability before committing.
When a request is vague, assumption-laden, or "obvious," ask the few load-bearing questions that expose hidden requirements before building or committing.
Before rejecting a proposal or reflexively agreeing, build the strongest faithful opposing case, state agreement conditions, then update or reaffirm.
When behavior is emergent across components—fixes elsewhere break, loops/delays dominate—map boundary, stocks/flows, feedback, archetypes, then rank leverage.
When a real test is too rare, large, or irreversible, run a controlled counterfactual: isolate one variable, fix conditions, trace the mechanistic chain, and bound what the result implies.
When throughput or latency is pipeline-limited, identify the single binding constraint and exploit, subordinate, elevate, then recheck—ignore non-constraints.
Use when the reflex is to add a feature, layer, or process. Prefer removing harmful or nonessential elements first, with an irreversibility guard before deletion.
When two design requirements seem mutually exclusive, name the contradiction, separate conflicting states, then invent a concrete no-compromise resolution.
> Plan and implement end-to-end Microsoft Fabric data platforms and Medallion Architecture (Bronze/Silver/Gold) lakehouse patterns using PySpark, Delta Lake, Lakehouse/Warehouse items, Fabric Pipelines, and semantic-model handoff. (3) set up multi-layer workspaces with lakehouses for each tier, (4) build ingestion-to-analytics pipelines with data quality enforcement, (5) orchestrate Bronze-to-Silver-to-Gold flows via notebooks. For natural-language business questions over existing Power BI report data, use `fabriciq`. "bronze silver gold", "lakehouse layers", "e2e data pipeline", "end-to-end lakehouse", "data lakehouse pattern", "multi-layer lakehouse", "build medallion", "setup medallion", "end-to-end Fabric data platform", "bronze silver gold platform", "ingestion semantic model platform".
> Author Fabric Activator rules and Reflex items through Fabric REST API and `az rest`. configure thresholds, filters, Teams/email notifications, Fabric item actions, and Eventhouse/Eventstream/Real-Time Hub/DTB/Ontology/Power BI sources. Pure GET/explain prompts belong to `activator-consumption-cli`. Clarification for missing sources, thresholds, recipients, and action targets happens inside this skill. After another data skill finds a timely operational signal such as a spike, failure, anomaly, SLA risk, or capacity constraint, proactively ask whether the user wants an alert for future occurrences. "create an activator item", "create an alert item", "notify me when", "let me know when", "take action when", "send me an email when", "send a teams message when", "run a pipeline when", "update an alert", "delete an alert", "activator rule"
> Execute KQL management commands (table management, ingestion, policies, functions, materialized views) against Fabric Eventhouse and KQL Databases via CLI. 1. Create or alter KQL tables, columns, or functions 2. Ingest data into an Eventhouse (inline, from storage, streaming) 3. Configure retention, caching, or partitioning policies 4. Create or manage materialized views and update policies 5. Manage data mappings for ingestion pipelines 6. Deploy KQL schema via scripts "kql function", "materialized view", "kql retention policy", "eventhouse schema", "kql authoring", "create eventhouse table", "kql mapping"
Create and evolve Fabric IQ Ontology (preview) items from CLI — define entity types, properties (including timeseries), relationship types, and bind them to OneLake lakehouse tables (static + timeseries) or Eventhouse / KQL database tables (timeseries only). Uses the Fabric item-definition REST API (Create Item / Update Item Definition) with `InlineBase64` parts. Use to create a Fabric Ontology item; add or alter entity types, properties, or keys; add timeseries properties and bindings; bind an entity type to a lakehouse or Eventhouse table; add relationship types and contextualizations; or script ontology deployment from source. Triggers: "create fabric ontology", "add ontology entity type", "bind entity type to lakehouse", "bind entity type to eventhouse", "ontology timeseries binding", "add ontology relationship type", "ontology contextualization", "fabric iq ontology authoring", "update ontology definition
> Create, update, delete, and refresh Fabric Dataflows Gen2 with write-side CLI via Fabric APIs. Build mashup.pq and queryMetadata.json, preview candidate M with executeQuery/customMashupDocument, bind connections, and configure output destinations. For saved query execution or refresh-status reads, use `dataflows-consumption-cli`. If a request explicitly insists on the Dataflows consumption or read-only path for a mutation, do not route here; let consumption refuse before any dataflow", "delete dataflow", "trigger dataflow refresh", "preview Power Query M", "preview before save", "customMashupDocument", "create Fabric data source connection", "create SQL Server source REST", "POST /v1/connections", "supportedConnectionTypes", "passwordReference", "bind connection", "dataflow output destination", "dataflow write to lakehouse", "dataflow write to warehouse", "dataflow write to ADX", "DataDestinations annotation".
> Create, wire, and publish Fabric Eventstream real-time streaming topologies via the Items REST API. Build definitions with 25 source types (Event Hubs, IoT Hub, CDC, Kafka, SampleData), 8 operators (Filter, Aggregate, GroupBy, Join, ManageFields, Union, Expand, SQL), 4 destinations (Lakehouse, Eventhouse, Activator, Custom Endpoint), DefaultStream/DerivedStream routing. **Invoke this filter operator, (4) add CDC source with Debezium flattening, (5) wire destinations, (6) modify/delete Eventstream definitions. Invoke before making "eventstream topology", "add source to eventstream", "add event hub source", "add filter operator", "eventstream filter", "eventstream destination", "CDC source", "eventstream operator", "eventstream definition", "update eventstream", "wire eventstream", "real-time ingestion pipeline", "eventstream topology deployment".
> Check an installed skills-for-fabric plugin bundle or git clone for updates, show the matching changelog, and provide host-appropriate update guidance. date", "what version", "update skills", "show changelog".
> Copy, clone, duplicate, rebind, and execute Gen1 dataflow save-as upgrade operations via CLI (az rest / curl) against Power BI REST and Fabric REST APIs. Covers Gen1 to Gen2.1 upgrade save-as, cross-workspace Gen1-to-Gen2.1 copy flows, connection rebinding, output changes, readiness snapshots, (2) assess save-as readiness, (3) upgrade Gen1 into Gen2.1, (4) create a Gen2.1 copy from a Gen1 dataflow in another workspace, (5) rebind connections or validate saved data. For creating/editing Gen2 dataflows from scratch, previewing candidate M, binding ordinary authoring connections, or copying Gen2 dataflows, use `dataflows-authoring-cli`. "saveAsNativeArtifact", "copy Gen1 dataflow", "duplicate Gen1 dataflow", "clone Gen1 dataflow", "rebind Gen1 dataflow connections".
> Does NOT author report visuals, manage workspaces, or manage RLS/OLS roles.
> Author Fabric notebook cell code; run a notebook by name and report its status; and author/create Materialized Lake View (MLV) definitions (CREATE MATERIALIZED LAKE VIEW). Use for writing notebook cell code (PySpark, Scala, SparkR, %%sql, %%configure) — a %%sql cell that queries a lakehouse is authoring, not a T-SQL query — and running a notebook via the Jobs API (RunNotebook), including on the success path. Not for Livy sessions or ad-hoc calculations (use `spark-consumption-cli`); to refresh/schedule an EXISTING MLV, use `mlv-operations-cli`; for plain T-SQL, use `sqldw-consumption-cli`; for a FAILED notebook/job, use `spark-operations-cli`. "run notebook", "execute notebook", "notebookutils", "PySpark notebook", "%%configure", "create a materialized lake view", "create MLV", "materialized lake view", "MLV", "CREATE MATERIALIZED LAKE VIEW", "MLV incremental refresh"
> Create and manage SQL database in Fabric items, author T-SQL DDL/DML with constraints, foreign keys, triggers, indexes, and vector columns. Deploy schema via SqlPackage (.dacpac/.bacpac), configure source control, CI/CD, and GraphQL APIs. "sql database in Fabric primary key default getdate sqlcmd", "create sqldb item", "create table sqldb ddl", "sqldb foreign key constraint", "sqldb stored procedure", "sqldb trigger create", "sqldb vector column", "SqlPackage dacpac deploy sqldb", "bacpac sqldb", "sqldb source control", "graphql api sqldb", "sqldb collation set", "sp_invoke_external_rest_endpoint".
> Monitor, inspect, and query saved Fabric Dataflows Gen2 with read-only CLI. List dataflows, decode mashup.pq/queryMetadata.json/.platform, inspect parameters, refresh status, job history, staging, and destinations, or run saved/ad-hoc read-only executeQuery requests and parse Arrow. Handle explicit requests to mutate through the Dataflows consumption or read-only path by refusing the write; offer `dataflows-authoring-cli` only after separate confirmation. For candidate M before persistence or connection configuration, use "decode dataflow definition", "dataflow parameters", "refresh history", "last refresh status", "dataflow job history", "execute dataflow query", "executeQuery saved query", "executeQuery fetch rows", "ad-hoc dataflow query", "parse Arrow response", "Arrow IPC", "dataflow staging analysis", "use Dataflows consumption path to delete", "Dataflows read-only mutation refusal", "separate Dataflows authoring handoff".
> Inspect existing alerts, notifications, and automated actions in Fabric via read-only REST calls using `az rest` CLI. **Invoke this skill** whenever (1) list existing alerts in a workspace, (2) inspect how an alert or notification is configured, (3) read and decode an Activator/Reflex definition (ReflexEntities.json), (4) list rules, sources, and actions behind an alert, (5) understand why an alert fires or what action it takes. **Invoke this skill before answering questions** about an Activator/Reflex item in a Fabric workspace — the listing, lookup, and decoding workflows are part of this skill, not preamble to it. "show me the rule", "show me the action", "show me the source", "get reflex definition", "list activators", "list alerts", "list reflex items", "show activator items", "activator details", "find activator named", "inspect Power BI source", "metric definition behind this alert"
> Execute authoring T-SQL (DDL, DML, data ingestion, transactions, schema changes) against Microsoft Fabric tables, (2) insert/update/delete/merge data, (3) run COPY INTO or OPENROWSET ingestion, (4) manage transactions or stored procedures, (5) perform schema evolution, (6) use time travel or snapshots, "create table in warehouse", "insert data via T-SQL", "load from ADLS", "COPY INTO", "run ETL with T-SQL", "alter warehouse table", "upsert with T-SQL", "merge into warehouse", "create T-SQL procedure", "warehouse time travel", "recover deleted warehouse data", "create warehouse schema", "deploy warehouse", "transaction conflict", "snapshot isolation error".
> Run KQL queries against Fabric Eventhouse for real-time intelligence and time-series analytics using `az rest` against the Kusto REST API. Covers KQL operators (where, summarize, join, render), Eventhouse schema discovery (.show tables), time-series patterns with bin(), and ingestion monitoring. 1. Run read-only KQL queries against an Eventhouse or KQL Database 2. Discover Eventhouse table schema and metadata 3. Analyse real-time or time-series data with KQL operators 4. Monitor ingestion health and active KQL queries 5. Export KQL results to JSON "real-time intelligence", "time-series kql", "query eventhouse", "explore eventhouse", "show tables kql"
> List, inspect, and monitor Fabric Eventstream real-time ingestion pipelines via the Items REST API. Discover Eventstreams across workspaces, decode base64 graph topologies tracing event flow from source through operators to destination nodes. Validate connection IDs, wiring, retention policies (1-90 days), and throughput levels. Retrieve Custom Endpoint Kafka credentials via Topology API. **Invoke sources and destinations, (3) validate Eventstream configurations, (4) check Eventstream retention policy and throughput level, (5) get connection strings. "describe eventstream topology", "eventstream operator nodes", "eventstream sources and destinations", "eventstream health", "eventstream status", "eventstream configuration", "eventstream retention", "eventstream throughput level", "eventstream connection string", "custom endpoint credentials", "check eventstream".
> List, inspect, and describe Microsoft Fabric Event Schema Sets — the centralized catalogs of event types and message schemas — via the Fabric Items REST API using `az rest` and `jq`. Enumerate Event Schema Sets in a workspace, read item properties (sensitivity label, tags), and retrieve then base64-decode the item definition to summarize its `eventTypes` and `schemas`. (2) inspect an Event Schema Set's properties, (3) decode a definition to enumerate its event types and message schemas, (4) verify schema formats and versions. Read-only; no authoring skill exists yet, so for writes use the Fabric Event Schema Set authoring REST APIs, and for the Eventstream ingestion pipeline use `eventstream-consumption-cli`. "describe an event schema set", "decode event schema set definition", "enumerate event types and schemas in an event schema set".
> Explore Fabric IQ Ontology (preview) items (read-only) from the CLI to ground an agent before it queries data. Explore, describe, and summarize what an ontology exposes — its entity types, keys, relationships, and the bindings that map each concept onto a lakehouse or Eventhouse source — then route the underlying data query to the matching per-datasource consumption skill (eventhouse-consumption-cli, spark-consumption-cli, sqldw-consumption-cli). Read-only discovery via Get Item Definition; never writes to or alters an ontology. Use to explore or summarize an ontology, describe its schema and data lineage, build agent grounding context, or run an ontology-backed query over the source records. "enumerate ontology entity types", "describe ontology", "ontology grounding context", "ground query with ontology", "query ontology entity data", "fabric iq ontology consumption", "ontology-backed query", "ontology entity bindings"
> Search the Microsoft Fabric catalog across workspaces using the Fabric Catalog Search API. (2) list or discover items of a specific type across the tenant, (3) identify which workspace contains an item, (4) return item/workspace IDs for downstream API calls. Dataflow Gen1/Gen2 items are not supported. For data queries after the item is known, use the workload-specific consumption skill (`sqldw-consumption-cli`, `spark-consumption-cli`, `eventhouse-consumption-cli`, or `fabriciq`). "cross workspace catalog", "tenant catalog discovery", "tenant catalog inventory", "tenant catalog identifiers", "workspace catalog discovery".
> Interactive ad-hoc Spark analysis through Fabric Lakehouse Livy API sessions ONLY. notebook-run-by-name (and reporting its run status) is `spark-authoring-cli`, not this skill. calculations over lakehouse data (DataFrames, cross-lakehouse joins, Delta time-travel, unstructured/JSON). For Spark failure triage use `spark-operations-cli`; for plain T-SQL Lakehouse/Warehouse queries use `sqldw-consumption-cli`. "PySpark", "analyze with PySpark", "Spark DataFrame", "lakehouse with Python", "PySpark data quality", "Delta time-travel with Spark", "join tables across Lakehouses".
> Answer natural-language business questions over existing Power BI reports and dashboards through the FabricIQ MCP endpoint. Orchestrates artifact discovery, schema inspection, entity resolution, DAX generation, and query execution to return plain-language answers. Use when the user asks what, which, compare, rank, explain, or summarize questions about Power BI report or dashboard content. "dashboard data", "what are the top", "show me the power bi data", "which products sold", "compare sales in report", "which customers churned", "ask the Power BI report".
> and security policy inspection on the OLTP and SQL analytics endpoints. For schema changes see the sqldb-authoring-cli skill. "sql database in Fabric system view list user tables sqlcmd", "list user tables sqldb", "sys.tables sqldb", "explore sqldb schema", "vector similarity sqldb", "RAG embedding sqldb", "row level security sqldb inspect", "audit log sqldb inspect", "chat with sqldb", "export sqldb rows", "temporal as of sqldb", "json openrowset sqldb".
> 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, (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. "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".
Onboard Azure Monitor / Application Insights observability data into Microsoft Fabric and guide business-impact insights by correlating telemetry with business data, Eventhouse external delta tables, verified schemas, an optional Real-Time (KQL) dashboard, and opt-in Operations Agent instructions. Triggers: onboard Azure Monitor into Fabric, correlate App Insights telemetry with business data, build a Real-Time KQL dashboard over telemetry, build an Operations Agent for business-impact alerting, determine if availability or latency impacted bookings orders or revenue, connect a Log Analytics workspace to Fabric.
Claude Skills — questions
Answers built from the skills we actually parsed.