Compress natural language memory files (CLAUDE.md, todos, preferences) into caveman format to save input tokens. Preserves all technical substance, code, URLs, and structure. Compressed version overwrites the original file. Human-readable backup saved as FILE.original.md. Trigger: /caveman-compress FILEPATH or "compress memory file"

Decision guide for delegating to caveman-style subagents. Tells the main thread WHEN to spawn `cavecrew-investigator` (locate code), `cavecrew-builder` (1-2 file edit), or `cavecrew-reviewer` (diff review) instead of doing the work inline or using vanilla `Explore`. Subagent output is caveman-compressed so the tool-result injected back into main context is ~60% smaller — main context lasts longer across long sessions. Trigger: "delegate to subagent", "use cavecrew", "spawn investigator/builder/reviewer", "save context", "compressed agent output".

Implementation & review guide for suspend/resume (revalidation) on query-engine iterators in src/redisearch_rs/rqe_iterators/. Use this when adding the Box<Self> suspend/resume mechanism to an RQE iterator that currently only implements the legacy revalidate, when modifying an existing suspend/resume implementation, or when reviewing a PR that touches this machinery (boxed.rs, ref_mode markers, child-slot helpers, layout-invariant proofs).

Create, configure, and manage Netlify deploys from code — reach for this when setting up Git continuous deployment, running netlify deploy or netlify deploy --prod from the CLI, writing netlify.toml deploy contexts, adding a Deploy to Netlify button, wiring build hooks, configuring Deploy Previews or branch deploys, locking or skipping deploys, fixing a failed or secrets-scanning deploy, or when someone asks to "deploy my site", "set up preview deploys", "add per-branch build config", "keep staging out of Google", or "add a deploy button to my README".

Set up a local Grove environment for running code example tests. Use when the user asks to "set up Grove", "configure the test suite", "get started with Grove", "set up my environment", "I need to run Grove tests", or is working with Grove for the first time and needs prerequisites configured. Checks for required tools, installs dependencies, guides .env file setup, verifies MongoDB connectivity, and checks sample data availability.

Create and schedule Threads posts, multi-part chains, image carousels and reply control via Publora MCP. Use when the user wants to post to Meta's Threads, including long content that Publora splits into a connected chain of replies past 500 characters. Not for Instagram (use instagram-post) or X threads (use x-post).

Use when working in a repository that uses bd or Beads for durable project task tracking, issue dependencies, blocker management, multi-session handoff, or shared work memory. Trigger when the user asks to find ready work, claim or close tasks, create follow-up work, inspect blockers, recover project context, or choose between local planning and persistent project tracking.

Receive and authenticate 360dialog WhatsApp Business API webhooks. Use when setting up a 360dialog webhook endpoint, adding HTTP Basic auth to the webhook headers, debugging 401s, or handling inbound messages, message status updates (sent, delivered, read, failed) and Coexistence events such as smb_message_echoes.

Create, review, test, rebase, and rollback Alembic database migrations for OPRE OPS. Use this skill whenever the user mentions database migrations, alembic, schema changes, adding/modifying columns or tables, model changes that need migration, "migrate the database", or a migration chain conflict / divergent heads after main has moved forward. Also use when a model change has been made and the user needs to generate the corresponding migration.

Investigates CPU, memory, and contention problems with Grafana Pyroscope profiles via the gcx CLI. Compares affected and baseline workloads, locates changed application stacks, and tests root-cause hypotheses against symptom and change evidence. Use when a service shows high CPU, growing memory, GC pressure, lock contention, or hot functions, when code regressed after a deploy, or when given a suspect service, flame graph, trace, or profiling anomaly report. For profiling setup use pyroscope; for profilecli-based analysis use profilecli-insights.

Converts and migrates legacy Orbit JSON DSL queries to Orbit GQL, then checks each converted query read-only against the graph. Legacy queries are objects with query_type, nodes, relationships, or aggregations. Use when a repo, runbook, script, doc, or agent skill contains them. Also use when a `glab orbit query --file` envelope or API body has an object `query`. Also use when a legacy JSON query fails with HTTP 400 compile_error. Do not use for writing new GQL queries (use the orbit skill).

Build, integrate, audit, or upgrade Stream Chat, Video, Feeds, and Moderation in React / Next.js web apps - the default skill for web work. Scaffold a new app, add Stream to an existing project, review an integration, upgrade an SDK major, or migrate from Sendbird. Covers team messaging, DMs, video conferencing, livestreaming, social feeds, AI assistants, and apps built on Stream Agents (voice and text agents, @stream-io/calypso). Web React only - React Native uses stream-react-native.

Use when the work-journal app asks to summarize one hour of recorded activity into a narrative, a timeline, detailed notes, a main focus, and action items.

Decide whether flow metrics would actually help your context before adopting any. Interviews you about the symptoms and friction you want to improve, then connects those symptoms to the flow metrics that would make the problem visible — WIP, age, throughput, cycle time, distribution — and ends in one small 2-4 week experiment. Use when a team or portfolio is considering flow metrics, when someone is measuring flow but cannot say what decision the numbers change, or when a metrics dashboard exists and nothing has improved. This decides *whether* to adopt metrics at all; to set actual WIP numbers once you have, use wip-limit-configuration-coach. If the question is broad — AI pilots with no clear return, or "where do we start" — run ai-activity-to-impact first; it finds the constraint and routes here.

Extend a board that ends at Released, Deployed or Done with one adoption/learn lane, so a team stops treating deployment as success. Coaches where to put the lane and at what altitude, its entry and exit criteria, what telemetry has to exist first, how to limit it, how to backfill the work already closed as Done, and how to answer the overhead objection honestly. Use when features ship and nobody knows whether anyone uses them, when AI use cases are deployed and unused, or when a spec-driven board has a column for every step of the software and none for the human on the other end. For the numbers behind the limit this puts on the new state, use wip-limit-configuration-coach. If the epics themselves are written as output, run outcome-framing-coach first. If the question is broad — AI pilots with no clear return, or "where do we start" — run ai-activity-to-impact first; it finds the constraint and routes here.

Use when you need to write unit tests for Micronaut applications — Mockito-first with @ExtendWith(MockitoExtension.class), @MicronautTest with @MockBean, HttpClient @Client(/) assertions, @Property overrides, @ParameterizedTest, and *Test vs *IT naming. For framework-agnostic Java use @131-java-testing-unit-testing. This should trigger for requests such as Add or improve unit tests in a Micronaut project; Reduce unnecessary @MicronautTest usage with Mockito-first tests; Write Mockito-first unit tests for Micronaut services; Mock Micronaut bean collaborators in unit tests; Review fast Micronaut tests without application context. Part of Plinth Toolkit

MLflow 3 GenAI agent evaluation. Use when writing mlflow.genai.evaluate() code, creating @scorer functions, using built-in scorers (Guidelines, Correctness, Safety, RetrievalGroundedness), building eval datasets from traces, setting up trace ingestion and production monitoring, aligning judges with MemAlign from domain expert feedback, or running optimize_prompts() with GEPA for automated prompt improvement.

Use when building, editing, validating, testing, or debugging an n8n workflow through the n8n-mcp MCP server — designing a flow, configuring a node, writing an expression or Code node, wiring credentials, or fixing one that misbehaves. The entry-point skill for the n8n-mcp-skills pack: it routes you to the right specialist skill, gives working knowledge of every n8n-mcp tool from turn one, and states the rules that keep workflows from breaking in production. Always consult it first on any n8n, workflow, node, or automation task — even a quick one-off, and even when the user names no skill — because n8n's surface drifts between versions and the specialist skills prevent silent failures.

Automate Agencyzoom tasks via Rube MCP (Composio). Always search tools first for current schemas.

Automate Adyntel tasks via Rube MCP (Composio). Always search tools first for current schemas.