Update to 33CRM

{
  "description": "Review the 33CRM lead funnel, lead sources, qualification/win/loss outcomes, pipeline-stage distribution, and recent lead records. Use when the user asks about lead generation quality, funnel health, conversion, sources, or leads needing operational attention.",
  "included_files": [
    {
      "relative_path": "agents/openai.yaml",
      "size_in_bytes": 445
    }
  ],
  "name": "lead-funnel-review",
  "skill_md_contents": "    ---\n    name: lead-funnel-review\n    description: Review the 33CRM lead funnel, lead sources, qualification/win/loss outcomes, pipeline-stage distribution, and recent lead records. Use when the user asks about lead generation quality, funnel health, conversion, sources, or leads needing operational attention.\n    ---\n\n    Use this skill for lead-funnel questions.\n\nWorkflow:\n1. Resolve the requested period.\n2. Call `sales_summary` first for lead-cohort totals, qualified/attended/won/lost metrics, by-stage and by-source breakdowns.\n3. Use `list_leads` when the user needs the underlying recent leads, workflow status, source, stage, or owner.\n4. Use `get_lead` only after identifying a specific lead and when stage history is needed.\n5. Use `pipeline_summary` only as supplemental current deal-pipeline context; do not present it as the lead funnel.\n\nOutput:\n- Funnel overview\n- Source mix\n- Stage / outcome mix\n- Leads needing follow-up based on returned workflow state when requested\n- Data quality / pagination notes\n- Clear distinction between lead funnel and deal pipeline\n\nRules:\n- Do not infer ad attribution from a lead source beyond what the CRM explicitly stores.\n- Do not infer reasons for won/lost outcomes from fields that are not returned.\n- Do not expose or imply clinical information.\n- Reply in the user's language.\n- Read-only only.\n"
}