program-scorer
Skill1 starsChanged 47 days agoMITView on GitHub
Audit any Klaviyo flow against a scoring rubric and produce a scored report with prioritized fixes.
In this repository9 other files it gives its agents.
- abandonment-recoverySkill
- brief-to-buildSkill
- post-purchaseSkill
- pre-send-qaSkill
- replenishmentSkill
- sunset-suppressionSkill
- _templateSkill
- welcome-seriesSkill
- winback-reactivationSkill
The file
Files
Program Scorer
Status: Planned. This is a stub. Section bodies are TODO markers. See
skills/_template/SKILL.mdfor what belongs in each section.
Scope: Cross-cutting flow audit and scoring.
This skill is read-only. It guides analysis, builds, audits, and QA. It never sends email or SMS.
When to use this skill
TODO: Trigger phrases and situations that should load this skill. List what it is NOT for and point to the sibling skill that covers that case.
Inputs
TODO: Required inputs versus optional inputs. Instruct the agent to ask for anything missing instead of guessing.
Program blueprint
TODO: Objective and audience. Message architecture (each touch: purpose, timing, channel, and the job it does). Timing as guidance with reasoning, not rigid rules. Splits and personalization. Exits and suppression. Content blocks described by the job each block does, not fixed copy.
Klaviyo build mapping
TODO: Flow trigger, profile filters, flow filters, conditional splits, and required catalog or event data. Starting point: Which flow settings and data the scorer reads to rate each rubric dimension. For tool names, arguments, and API shapes, defer to klaviyo-labs/agent-context and the live Klaviyo API or MCP catalog. Never trust this skill over the live catalog.
QA
TODO: What "ready to ship" means for this program, as a checklist. The skill guides QA and never sends.
Measurement
TODO: Primary KPI, guardrail metrics, attribution approach, and holdout guidance. Label every benchmark or example number illustrative unless it cites a verifiable source.
Worked example
TODO: Add examples/example-brief.md with a brief-to-build-spec walkthrough for this program, then point to it here.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

