Product collection · 25 skills
25 skills for connectors, pipelines, databases, analytics platforms and other data products.
Work from an approved sample to correct real-data output, reconciliation, recurring workloads and production approval. These versions account for access, source/destination behavior, schemas, freshness, recovery, governance and workload economics.
Each version has its own skill name, product-specific decision criteria and examples. Shared methods do not create a runtime dependency on the other repository.
Clone jonbishop1/data-product-marketing-skills and copy the desired .agents/skills/<name>/ folders to the location documented by your Agent Skills-compatible host. Preserve the references and assets.
For Claude Code: /plugin marketplace add jonbishop1/data-product-marketing-skills then /plugin install data-product-marketing-skills@jon-bishop-data-product.
The skills use the open Agent Skills format and MIT license. File-format and plugin-manifest checks have passed; compatibility and output quality still depend on your host, tools and evidence.
Examples are fictional unless explicitly described as measured evidence. A sample result does not establish production readiness. Plans and drafts do not authorize external actions.
Design a bounded workload evaluation that checks correctness, reconciliation, recovery, freshness and cost.
Agree on the decision, baseline, representative authorized workload, correctness invariants, updates/deletes/late arrivals, recovery, freshness, cost cap, ownership, and cleanup before results. Keep proposed criteria and measured outcomes distinct.
Assess whether an agent can discover, understand and complete a defined task, with evidence for each stage.
Assess a defined data task through documented interfaces, scoped access, schema discovery, query/workload cost, correctness, freshness, and partial-failure recovery. Public-only evidence cannot prove authenticated execution or production data readiness.
Prepare useful technical contributions for relevant communities, maintainers and product ecosystems.
Choose source/warehouse ecosystems and practitioner communities where the workload is relevant. Contribute a synthetic reconciliation example, operational lesson, or transparent methodology; never expose private customer queries or records.
Build a small, verifiable path to a useful technical result, including prerequisites, recovery and cleanup.
Use synthetic or authorized records with explicit schema, expected keys/values, reconciliation checks, connector permissions, resource caps, and cleanup. A green connection or sample dashboard alone is not the success assertion.
Work out who needs your product, how they describe the problem, and where you can reach them.
Compare segments by source-to-consumer workflow, correctness/freshness need, access authority, and operational ownership. Distinguish analyst interest, engineering evaluation, platform review, and budget authority.
Compare the alternatives buyers consider, including their current manual process, and find a useful position for your product.
Compare the workload against SQL jobs, open-source connectors, and managed platforms. Require equivalent update/delete semantics, reconciliation, freshness, failure recovery, access model, and cost conditions before claiming superiority.
Turn a clear brief and real expertise into a useful article, guide, comparison or other content asset.
For guides state data provenance, shape, transformations, expected invariants, access scope, reconciliation, and operating limits. Label synthetic records and illustrative charts; do not publish a performance conclusion without equivalent, reproducible workloads.
Decide which customer questions your content should answer and where that content belongs.
Build around source/destination behavior, schema changes, late arrivals, reconciliation, freshness, governance, and workload cost. Require a verified example or defensible evidence for each proposed connector or performance topic.
Turn your product facts and customer language into clearer website, email, ad or sales copy.
Anchor copy in the data workflow and the output users can validate. Qualify freshness, correctness, scale, and cost claims; distinguish connectivity from correct recurring operation. Offer a sample or approved workload evaluation with clear prerequisites.
Use customer evidence to understand buying decisions, switching concerns and friction in your marketing or product experience.
Investigate concern about silent corruption, credential scope, data movement, migration, or operating a failed workload. Offer reconciliation evidence and an approved bounded evaluation; retain governance controls rather than treating them as persuasion problems.
Connect marketing activity to signup, activation, retention and customer quality using clear definitions and usable data.
Choose workload/workspace/account identities deliberately. Separate sample runs, valid real-data output, reconciliation, recurring approved operation, and paid adoption. Include empty/failed runs, freshness, operating cost, and complete observation windows.
Plan or diagnose paid search around buyer intent, campaign structure and meaningful conversion tracking.
Tie query groups to supported source/destination workloads and buying context. Qualify access, platform fit, correctness needs, and an evaluation owner; account for delayed enterprise qualification and distinguish students from deployable accounts.
Plan a page around the visitor’s questions, the evidence they need and the action you want them to take.
Organize the page around workload, source-to-destination flow, supported semantics, reconciliation evidence, permissions, cost, and evaluation. Separate credential-free sample access from approved real-data setup instead of promising instant production access.
Create a reusable product, audience and adoption brief. Other skills can reuse it, but a full intake is optional.
Record source and destination, data sensitivity, update patterns, schema, freshness/correctness requirements, access owner, and workload economics. Define reconciled real-data output and approved recurring operation separately.
Choose marketing priorities that fit your product, audience, budget and capacity.
Sequence workload-specific demand, access approval, representative correctness evaluation, and operating ownership. If credentials or reconciliation block progress, resolve that path before increasing demos. Measure approved real-data evaluations separately from sample runs.
Plan or review Facebook and Instagram campaigns with attention to creative, audience, tracking and downstream quality.
Judge whether a synthetic workflow explanation or proof asset reaches a credible evaluator. A content download does not establish platform authority or approved data access. Preserve data privacy and evaluate downstream workload fit.
Help new users reach a useful first result through clearer signup, setup, in-product guidance and supporting messages.
Map approved access, source connection, initial output, reconciliation, incremental changes, recurring runs, and production approval. Keep sample success and real-data success separate; show the next useful step while an administrator reviews access.
Prepare the launch of a product or feature, from audience and channel choices through follow-up.
Verify connector behavior, approved evaluation data, reconciliation, access scope, cost limits, operating ownership, and recovery before choosing launch scope. A sample-only test can support a scoped beta claim, not unqualified production availability.
Diagnose technical, on-page and content problems, then prioritize fixes using evidence from your own site.
Inspect connector and workload pages for distinct supported semantics, source evidence, evaluation instructions, and appropriate canonical structure. Check whether thin integration permutations obscure the useful documentation and proof.
Find relevant search topics, understand search intent and choose opportunities your site can realistically pursue.
Separate conceptual education from operational source/destination, schema, recovery, freshness, and cost queries. A connector page needs verified behavior and a distinct task; do not map every possible product pair to an indexable page.
Turn useful event conversations into consent-aware, qualified next steps and bounded evaluations.
Use synthetic records and a workload-evaluation worksheet. Qualify source/destination, correctness/freshness need, access owner, and platform approval without collecting sensitive data. Track agreed evaluations separately from scans and conversations.
Explain and distribute a verified joint workflow, with clear compatibility, data flow and support ownership.
Verify source, transformation, destination, schema/update semantics, credentials, reconciliation, freshness, partial failures, and recovery. Connectivity alone does not prove correct operation; distinguish a tested supported connector from a custom example.
Choose a value metric and package boundaries that fit real technical workloads and predictable customer costs.
Compare meters against volume, scan/compute behavior, concurrency, refresh frequency, retention, egress, and operating cost. Explain metering semantics, retries, failed jobs, predictability, and spend controls with explicitly observed or illustrative workloads.
Substantiate technical claims with an appropriate test, reproducible evidence and honest limitations.
Prove equivalent output and semantics before comparing workload speed or cost. Record dataset shape and skew, updates, cache state, concurrency, freshness, failures, resources, repeated runs, and uncertainty. Include unfavorable conditions and limitations.
Help a technical champion turn demonstrated value and unresolved risks into an informed internal decision.
Turn a reconciled evaluation slice into an honest decision memo: source/destination fit, approved data boundary, recovery, recurring operation, procurement, cost predictability, and owner. State remaining production tests and a no-go option.