name: emaildeliverabilityskill description: "Use when Codex, Hermes, OpenClaw, Claude Code, Cowork, or another AI agent needs to plan, review, implement, audit, or improve email work focused on inbox placement, authentication, sender reputation, complaint control, and remediation workflows. Triggers include requests about DNS authentication audits, warmup plans, complaint spike triage, blocklist checks, bounce analysis, and reputation recovery."

Email Deliverability Skill

Protect the sender before chasing growth. Separate authentication, reputation, content, list quality, and provider issues before recommending action.

Scope

inbox placement, authentication, sender reputation, complaint control, and remediation workflows.

Request Signals

DNS authentication audits, warmup plans, complaint spike triage, blocklist checks, bounce analysis, and reputation recovery.

Guardrails

1. Primary lens: inbox placement, authentication, sender reputation, complaint control, and remediation workflows.
2. Common request signals: DNS authentication audits, warmup plans, complaint spike triage, blocklist checks, bounce analysis, and reputation recovery.
3. Default posture: Protect the sender before chasing growth. Separate authentication, reputation, content, list quality, and provider issues before recommending action.
4. Useful output family: DNS/authentication audits, reputation triage notes, warmup schedules, suppression reviews, monitoring plans, and remediation runbooks.
5. Production boundary: separate recommendation from execution.
6. Evidence boundary: say which source material supports the recommendation.
7. Review boundary: identify the human owner for risky changes.
8. Data boundary: do not assume missing fields, consent, or suppression state.

Execution Path

1. Classify the deliverability problem: authentication failure, bounce spike, complaint spike, spam-folder placement, throttling, blocklist, or engagement decay.
2. Collect evidence by domain, provider, segment, campaign, and time window rather than treating all mail as one stream.
3. Check SPF, DKIM, DMARC, custom tracking domain, bounce domain, alignment, and recent DNS changes.
4. Review list acquisition, consent, suppression, frequency, reactivation, and recent imports for reputation risk.
5. Recommend a staged remediation plan with stop-loss thresholds, monitoring cadence, and owner for each action.
6. Keep DNS edits, suppression changes, imports, provider migrations, and production throttling behind explicit approval.

Reviewer Notes

• Authentication advice distinguishes missing records from alignment failures.
• Warmup or ramp guidance is tied to observed volume, complaint, bounce, and engagement patterns.
• Root-cause analysis avoids blaming content when list quality or authentication is the actual issue.

  访问小葱技能站7w4.net，解锁更多实用的AI技能插件。

• Remediation includes what to pause, what to monitor, and when to resume.
• The plan respects consent, unsubscribe, and complaint handling obligations.

Output Pattern

Return DNS/authentication audits, reputation triage notes, warmup schedules, suppression reviews, monitoring plans, and remediation runbooks. Keep recommendations concrete. Separate analysis from live-system actions, and require explicit approval before sending email, importing contacts, changing DNS, altering suppression rules, or editing production automations.

这个 Skill 质量中等偏上，文档清晰易懂，风险控制做得较好，能帮助识别高风险操作并要求审批。但实用性有限，缺少示例和具体操作指导，仅提供思路框架而没有可直接使用的工具或脚本，适合作为规划参考而非日常使用。总体来说概念清晰但落地能力不足。