Overview
In this article, you'll find a reference for all available Cordial skills for Claude — pre-built, downloadable skill bundles that give Claude the context and tool-use patterns it needs to complete specific marketing and platform tasks in your Cordial account.
Each skill is a self-contained bundle you install in Claude Desktop. Once installed, Claude knows exactly which Cordial MCP tools to call, in what order, and how to interpret the results — no prompt engineering required on your end.
Before you begin
To use Cordial skills, you'll need:
- Claude Desktop installed and running
- The Cordial Production MCP connected (
mcp.cordial.io) - Access to the Cordial Agentic Cookbook on GitHub
Note: Each skill is account-aware. Most skills require your Cordial account to be identified before Claude can query live data. You can establish account context by telling Claude which account to work with at the start of a conversation.
Install a skill
- Go to the Cordial Agentic Cookbook on GitHub.
- Download the
.zipfile for the skill you want to use. - Open Claude Desktop and navigate to Settings > Skills.
- Upload the
.zipfile as a new skill. - Start a new conversation, ensure your Cordial Production MCP is connected, and prompt Claude using the example phrases listed in each skill's section below.
Available skills
Audience AI Breakdown
Creates a report showing how contacts are distributed across Cordial AI attributes — such as purchase propensity, price sensitivity, engagement momentum, or send-time preference.
Example prompts
- "Break down my active email audience by AI attributes."
- "What percentage of my subscribers are high purchase-propensity?"
- "Profile my loyalty segment by crdl_ai dimensions."
Download: audience-ai-breakdown.zip
Audience Count Snapshot
Provides a live count of contacts matching a saved audience or custom criteria — without opening Audience Builder.
Example prompts
- "How many contacts are in my winback audience right now?"
- "What's the current count for contacts who purchased in the last 30 days?"
- "Give me a live audience count for [audience name]."
Download: audience-count-snapshot.zip
Audience Overlap
Analyzes overlap and differences between two audiences, including intersection size, members in one but not the other, and rule-by-rule comparison.
Example prompts
- "What's the overlap between my SMS subscribers and my email list?"
- "How many contacts are in audience A but not audience B?"
- "Diff these two audiences and explain why the counts differ."
Download: audience-overlap-diff.zip
Capability Adoption Audit
Audits which Cordial AI features are enabled in an account, which are actively being used, and which are available to activate — useful for QBRs and feature reviews.
Example prompts
- "What AI models and features do we have enabled?"
- "Which capabilities are we paying for but not using?"
- "Run a capability and adoption review for this account."
Download: capability-adoption-audit.zip
Contact 360 Trace
Explains what happened to a specific contact — confirming their record, consent state, suppression status, and send history — to produce a root-cause summary for anomalies or complaints.
Example prompts
- "What happened to this contact? Why didn't they receive our last send?"
- "Trace this contact's history and explain why they have four emails in a row."
- "Is this email address still subscribed, and when did it opt out?"
Download: contact-360-trace.zip
Data Dependency Trace
Traces a contact attribute back to its originating data job, supplement, and import feed to isolate the point of failure when a field goes blank, zero, or wrong.
Example prompts
- "Why is this attribute suddenly showing 0 for all contacts?"
- "What data job writes this field, and did the last import succeed?"
- "Trace this attribute back to its feed."
Download: data-dependency-trace.zip
Experiment Report
Determines the top-performing test variant and reports performance with statistical confidence — covering open rate, click rate, conversion, and revenue per email, with bot-adjusted metrics where applicable.
Example prompts
- "Who won the A/B test on this message?"
- "Show me subject-line test results with adjusted open rates."
- "Aggregate the split-test results across all sends of this automation."
Download: experiment-report.zip
Frequency Volume Revenue
Analyzes send volume and attributed revenue trends by engagement tier and time period to assess whether frequency changes are helping or hurting program performance.
Example prompts
- "Are we over-mailing? Show me send frequency vs. revenue."
- "Volume is down but revenue is up — show me the trend."
- "Break down sends per contact this period vs. last, by engagement tier."
Download: frequency-volume-revenue.zip
List Growth and Churn Trend
Breaks down monthly subscribes, unsubscribes, and net churn over a multi-year window — going beyond the platform's native 12-month dashboard limit.
Example prompts
- "Show subscriber growth by month for the last two years."
- "What's our net subscriber gain or loss per month, going back further than the dashboard allows?"
- "Monthly list size over time for our email channel."
Download: list-growth-churn-trend.zip
Orchestration Performance Report
Builds an accurate performance report for a Cordial orchestration — chaining node discovery to child sends to message-level performance so the numbers tie out correctly.
Example prompts
- "How is my welcome series performing?"
- "Show me results by node for this automation over the last 90 days."
- "Build a performance report for [orchestration name]."
Download: orchestration-performance-report.zip
Personalization Discovery
Explores what contact attributes exist in an account, what values they hold, and what data coverage looks like — so you know what's actually available to personalize on or segment by.
Example prompts
- "What can I personalize on in this account?"
- "What attributes exist and what values do they have?"
- "What's the data coverage for loyalty tier across my contacts?"
Download: personalization-discovery.zip
Program Performance Report
Builds a QBR-ready report of program performance across sends, engagement, and revenue — segmented by channel, promotional vs. transactional, and batch vs. automated — with period-over-period comparisons.
Example prompts
- "Give me an exec-level report on our program performance this quarter."
- "How is email vs. SMS performing this month vs. last?"
- "Break down sends and revenue by channel for the QBR."
Download: program-performance-report.zip
Send Health Diagnostic
Diagnoses why a send under-delivered, stalled, or shows an implausible delivered rate — reconciling intended audience size against queued, sent, delivered, and failed counts.
Example prompts
- "Why did this send only go to 3,000 of my 7,000-contact audience?"
- "This message is stuck processing — what's going on?"
- "Is the delivery rate on this send real, or is it inflated by failures?"
Download: send-health-diagnostic.zip
Tag Coverage Audit
checks how consistently your messages are tagged across dimensions like region, promo type, or channel — so you know before building a sliced report whether the data is clean enough to trust.
Example prompts
- "How many sends are untagged?"
- "What share of our sends are missing promo-type tags?"
- "Audit our message tagging before I build the QBR."
Download: tag-coverage-audit.zip
Where Used Lookup
Maps all downstream dependencies for a Sculpt block, contact attribute, coupon, or asset before it's edited or retired — so you know what will break before you make the change.
Example prompts
- "Where is this Sculpt block being used?"
- "What messages or orchestrations reference this attribute?"
- "Check all dependencies for this asset before I retire it."
Download: where-used-lookup.zip
Comments
0 comments
Please sign in to leave a comment.