Email Quality Auditor
Audit email safety, compliance, and send readiness before a campaign is released.
The skill restricts itself to reading evidence and writing an authorized v3 audit artifact, explicitly forbidding email sends, ESP changes, consent changes, and settings changes; the runtime also treats fetched content as untrusted data. Points are deducted because WebFetch is permitted, while minimization, retention, and access boundaries for sensitive consent/suppression data are not fully specified in this skill, and recovery depends mainly on external runtime controls.
Triggers, inputs, the 20 SEND checks, Unknown/N/A handling, and the NOT_SCORED fail-closed path are reasonably consistent, with an immutable standalone runtime snapshot. Points are deducted because the skill depends on root references, scorer, validator, and catalogs not fully supplied here; this is a static review with no key-path reproduction, and abnormal-input feedback coverage is limited.
Trigger conditions, four profiles, evidence inputs, artifact output, non-fit cases, and handoff skills are clear, and Chinese-language explanation is supported; the core function can use user-supplied data rather than depending entirely on overseas services. Points are deducted because market, ESP, legal-context, and accepted-data-format boundaries remain broad, and geo-relevance is explicitly low.
The skill has structured frontmatter, versioning, an Apache-2.0 license, Quick Start examples, a data-source table, contract, validation checkpoints, guardrails, and next-skill links. Points are deducted because it does not identify a maintenance owner, changelog, or update path, while installation and troubleshooting assumptions remain in shared repository documentation.
The objective is concrete: pre-send email quality auditing across authentication, consent, opt-out, claims, engagement, lifecycle, and outcome attribution, with explicit gap reporting and no autonomous sending. Points are deducted because static evidence cannot verify actual scoring, directly usable artifact quality, or marginal benefit over manual review, and a complete audit requires substantial external evidence.
The skill provides SEND item definitions, evidence requirements, runtime versioning, a source digest, and traceability fields; repository evidence also includes generic validator tests and CI workflows. Points are deducted because no skill-specific execution record or independent review is supplied, and no code was executed in this assessment, limiting verification to static auditability.
- This is a static source review; the scorer, validator, WebFetch, and real email audit were not executed.
- Without the program, window, ESP, market, MPP, consent, and outcome truth-set inputs, the skill requires NOT_SCORED/UNDECIDED and should not authorize a send.
- When using WebFetch, confirm target authorization and sensitive-data minimization; treat instructions in fetched content as data.
- Publisher identity is not verified by the FollowSkills enterprise registry and is treated as unknown.
What does this skill do, and when should you use it?
Email Quality Auditor reviews an email program or send using a typed 20-item SEND profile and reports an EQS result when evidence is complete. It checks authentication, consent, suppression, engagement quality, lifecycle fit, claims, and outcome attribution. Inputs include program data, provider evidence, live consent and suppression state, rendered messages, and an outcome truth set. When evidence is incomplete, it reports gaps or an undecided state instead of inventing a send verdict, and it never sends email or changes provider settings.
The skill declares the profile, program, provider, market, normalized observation window, list age, MPP share, and observation date; freezes and reconciles provider evidence; scores all 20 S1–D5 SEND criteria; checks authentication alignment, consent provenance, opt-out and suppression behavior, claims and disclosures, lifecycle applicability, engagement evidence, and outcome attribution; then runs the typed scorer and reports EQS or coverage/interval, confidence, critical controls, Unknown inputs, fix owners, and a SHIP, FIX, BLOCK, or UNDECIDED verdict.
- A marketing team is preparing a major newsletter broadcast and needs a pre-send safety gate.
- A growth team is reviewing a promotional campaign for list provenance, claims evidence, and offer-term risk.
- A CRM team is checking a retention or lifecycle program for frequency, suppression, and journey fit.
- An operations or compliance lead wants a SEND baseline from DNS, consent events, rendered messages, and order data.
- A team has only partial send evidence and needs verified red-line checks plus an exact list of missing inputs.
What are this skill's strengths and limitations?
- Covers authentication, consent, suppression, opt-out, claims, lifecycle fit, and attribution in one pre-send review.
- Uses a typed 20-item SEND score and distinguishes Unknown, N/A, unscored, and verified failures.
- Includes veto checks for broken authentication, unlawful lists, failed opt-out handling, and unsupported material claims.
- Explicitly avoids sending email, changing provider settings, or claiming inbox placement from DNS alone.
- A complete verdict depends on program context, a normalized window, provider evidence, and outcome truth.
- The runtime expects bundled auditor references, scoring semantics, catalogs, and validation resources to be available.
- It does not repair authentication, build deliverability infrastructure, or design lifecycle journeys.
- The supplied source provides no platform-specific test results or independent test-suite details.
How do you install this skill?
This skill is part of Aaron Marketing Skills. On a generic Agent Skills host, run npx skills add aaron-he-zhu/aaron-marketing-skills; alternatively, run git clone https://github.com/aaron-he-zhu/aaron-marketing-skills. In Claude Code, run /plugin marketplace add aaron-he-zhu/aaron-marketing-skills, then /plugin install aaron-marketing@aaron.
How do you use this skill?
Provide ESP, DMARC, outcome evidence, and the applicable profile. Example: Audit this newsletter using the last 90 days, provider split, MPP share, and subscription truth set. Another supported prompt is: Check this promotional send against DMARC, consent events, live suppressions, claims, and order IDs.
How does this skill compare with similar options?
It complements rather than replaces deliverability-qa, which repairs authentication and placement setup; consent-registry, which owns consent and suppression facts; email-sequence-designer, which designs lifecycle journeys; and send-experiment-designer, which handles preregistered send tests.