Explore AI Agent Skills & Claude Prompts

Apply eventually consistent storage correctly - reason about the inconsistency window, implement read your own writes (RYOWs), tune consistency via N/W/R quorums, use sloppy quorums + hinted handoff, reconcile via anti-entropy / Merkle trees, and resolve conflicts with last-writer-wins (and its data-loss caveat), version vectors / siblings, or CRDTs. Use when designing on Cassandra / DynamoDB / Riak / Voldemort / Cosmos DB, diagnosing stale reads or lost updates, or choosing a conflict-resolution strategy. Triggers - "stale reads", "tunable consistency", "N W R quorum", "RYOWs", "version vector", "CRDT", "last writer wins", "sloppy quorum", "hinted handoff", "Merkle tree anti-entropy", "siblings". Produces a consistency design with explicit N/W/R, conflict-resolution strategy, and operational guarantees.

data-storage-and-modeling-patterns

Apply data storage and modeling patterns: cache hierarchy, consistency paradigms (strong vs eventual), file/object/block storage, warehouse vs lake vs lakehouse, ingestion patterns (batch, streaming, CDC, snapshot vs differential), schema-on-write vs schema-on-read, dimensional modeling (Kimball star schema, Inmon 3NF, Data Vault), Slowly Changing Dimensions (SCD types 1/2/3), and distributed-query patterns (broadcast vs shuffle hash join, MapReduce). Use when designing storage layers, modeling a warehouse, choosing ingestion frequency, or evaluating a transformation approach. Triggers: "warehouse vs lake", "Kimball vs Inmon vs Data Vault", "SCD type 2", "schema on read", "CDC vs scheduled extract", "broadcast join", "data lakehouse", "Iceberg / Delta / Hudi". Produces a chosen storage architecture + data model with rationale.

consensus-and-strong-consistency

Apply strong consistency mechanisms in distributed systems - ACID transactions, two-phase commit (2PC) for cross-partition transactions and its blocking pathology, consensus algorithms (Paxos, Multi-Paxos, Raft) with terms / AppendEntries / heartbeats / quorum-based commit, linearizability vs serializability, TrueTime + commit-wait (Spanner), single-threaded per-partition execution (VoltDB), and deterministic transaction execution (Calvin / FaunaDB). Use when designing on Spanner / CockroachDB / VoltDB / etcd, picking a consensus algorithm, or diagnosing transactional behavior in a distributed DB. Triggers - "2PC", "Paxos vs Raft", "linearizability vs serializability", "TrueTime", "Spanner", "CockroachDB", "VoltDB", "etcd Raft", "ACID across partitions", "compensating action", "XA transaction". Produces a strong-consistency design with explicit algorithm choice and trade-offs.

mdm-and-federated-data-governance

Apply Master Data Management (MDM) styles (Consolidation, Registry, Centralized, Coexistence), federated governance via data contracts and policy automation, data catalog + metalake architecture, knowledge graphs for metadata, semantic layers, and access control models (ACL, RBAC, ABAC + PEP/PDP/PIP/PAP). Use when scoping MDM, choosing an MDM style, designing a data catalog, building governance automation, defining data contracts, or implementing fine-grained access control on data products. Triggers: "MDM strategy", "consolidation vs registry vs centralized vs coexistence", "data contract", "data catalog", "knowledge graph for metadata", "ABAC for data", "semantic layer for governance", "metalake". Produces a chosen MDM style + governance architecture with policy automation.

winning-by-design-sales-excellence-framework

Design GTM playbooks, customer journeys, and messaging using the Winning by Design framework. Use this skill when Alex says "WbD this", "design the playbook for [segment]", "map the customer journey for [product]", "SPICED the discovery call", "bowtie model for [motion]", or "messaging house for [offering]". Triggers on GTM design work — net-new motions, stalled playbooks, segmentation redesigns, or messaging refreshes. Applies WbD primitives (SPICED, bowtie, impact/effort, moments-that-matter) consistently across deliverables.

embedding-models-and-domain-adaptation

Choose, train, and adapt text embedding models for production. Covers SBERT vs cross-encoder tradeoffs, contrastive learning, domain adaptation via continued pretraining, few-shot classification with SetFit, and hard-negative mining. Use when building semantic search, RAG retrieval, classification with limited labels, when generic embeddings underperform on your domain, or when picking between cross-encoder and bi-encoder architectures. Triggers: "choose an embedding model", "domain-adapt our embeddings", "few-shot classification", "SetFit for X", "cross-encoder vs bi-encoder", "embedding fine-tuning", "hard negatives". Produces concrete embedding strategy.

ai-build-vs-buy-and-model-adaptation

Make strategic AI decisions: build vs buy vs hybrid, fine-tuning vs RAG vs grounding, synthetic vs real-world data. Use when scoping a new AI capability and deciding whether to develop in-house or use a third-party model, when choosing how to adapt a foundation model to your domain, when planning a data strategy where labeled real-world data is scarce, or when evaluating whether AI is even the right tool for a given problem. Triggers: "should we build or buy this?", "fine-tune or use RAG?", "do we need fine-tuning here?", "should we use synthetic data?", "is AI actually the right tool here?", "is this an Innovator's Dilemma situation?". Produces a structured decision rationale covering core-competency assessment, resource check, model adaptation choice, and data strategy.

content-correspondent

Post-event content and outreach sequencing for NYC AI/tech events. Use this skill any time Alex mentions attending an event, returning from an event, or wants to follow up after one — even casually ("just got back from a founder session", "I'm at PMF x AI tonight", "need to write something about last night"). Also triggers for: drafting LinkedIn posts from event observations, classifying contacts from an event, writing follow-up DMs after meeting someone, building a content sequence from field notes, or converting Wispr/Granola notes into outreach and posts. Also use when Alex asks about post-event strategy or how to turn room conversations into content. For reviewing how posts or outreach performed after the fact, use the event-signal-tracker skill instead.

schedule Updated 17 days ago

Apply Jobs-to-be-Done (JTBD) theory and the canonical interview techniques: Jobs Interviews, Switch Interviews (Bob Moesta), Inverted Switch (cancellation), Customer Case Research (CCR), the Four Forces of Progress (Push, Pull, Anxiety, Habit), the Switch Timeline (six phases), the Milkshake Marketing reframe. Use when scoping customer research, designing interview protocols, understanding why customers switch (or don't), uncovering unarticulated needs, separating jobs from solutions. Triggers: "JTBD interviews", "Switch technique", "Bob Moesta", "why did they switch", "cancellation interview", "Four Forces of Progress", "milkshake reframe", "jobs vs solutions", "uncover unarticulated needs". Produces an interview protocol + analysis approach.

the-challenger-sale

Run deals, pipeline, and messaging through The Challenger Sale framework — Teach, Tailor, Take Control; Commercial Teaching 6-step pitch; the 4 Rules; SAFE-BOLD audit; Mobilizer identification; Take Control scripts. Use this skill when Alex says "Challenger this", "Challenger pitch for [account]", "reframe the [problem] for [persona]", "commercial insight for [vertical]", "who's the Mobilizer on [account]", "Mobilizer check on [contact]", "am I talking to a Talker", "teach-tailor-take-control on [deal]", "build the reframe slide", "rational drowning on [buyer]", "I'm getting along too well with [contact]", "I need to challenge this buyer's thinking", "is my pitch BOLD or SAFE", "SAFE-BOLD audit", "what's the insight that leads to us", "foil check", "my champion is a Talker", "how do I take control of [deal]", "teaching pitch for [exec]", "score my reframe", or "rebuild the pitch". Triggers when deals are pleasant but not progressing, when the buyer is comfortable in status quo, when category education is requ