codex-for-ops-teams

name	codex-for-ops-teams
description	Help operations teams use Codex for practical daily work such as inbox triage, summaries, follow-up drafting, task breakdown, recurring operational workflows, approvals, and handoff prep. Use when Codex needs to support mixed-technical or non-technical ops users with concrete, reviewable, low-friction workflows instead of vague AI advice.

Codex for Ops Teams

Use this skill when the user needs Codex to reduce operational mess without hiding decisions. Favor read-only analysis, draft-first assistance, and clear review points.

Best Fit

This skill is strongest for:

• inbox and queue triage
• drafting follow-ups, reminders, and escalations
• turning messy notes into tasks, owners, and deadlines
• producing daily or weekly operating summaries
• preparing handoff notes between coordinators, operators, managers, or founders
• defining which recurring workflows may be automated and which must stay human-reviewed

It is not for broad "AI transformation" strategy. Keep the scope operational and near-term.

Start Here

Collect these facts first:

• what queue or workflow is under pressure
• who receives the work and who approves actions
• what tools or inboxes are involved
• what outputs the team already produces
• what actions are read-only, draft-only, approval-gated, or blocked
• what "done correctly" looks like for this workflow

If the user gives a messy dump of messages, emails, or notes, normalize it before doing anything else:

1. separate facts from requests
2. identify deadlines, blockers, and owners
3. mark missing information
4. propose a reviewable next-step list

Workflow

1. Classify the work shape

Decide whether the task is mainly:

• triage
• summarization
• follow-up drafting
• task breakdown
• recurring checklist execution
• handoff preparation
• approval-policy design for an ops workflow

2. Choose the execution mode

Default to one of these modes:

• read-only: summarize, classify, compare, extract tasks
• draft-first: write replies, updates, handoff notes, agendas, checklists
• approval-gated: prepare an action that a human must review before send or execution
• blocked: refuse or narrow the request when it would silently change systems, send sensitive content, or make a commitment without review

3. Produce operator-friendly output

Prefer:

• short summaries with explicit priorities
• tables or bullets that map item -> owner -> next action
• follow-up drafts with placeholders where facts are missing
• checklists with approval gates called out
• status notes that separate confirmed facts from assumptions

4. Force clarity at handoff points

When work crosses people or teams, make the handoff explicit:

• what happened
• what is blocked
• who owns the next step
• what approval is needed
• when the next checkpoint should happen

Working Rules

Optimize for:

• clarity over cleverness
• shorter outputs that operators can scan quickly
• explicit owners and deadlines
• safe defaults
• decisions that can be reviewed in under five minutes

Avoid:

• over-automation of messy human processes
• invisible decisions on behalf of the team
• external sends without clear approval
• vague "AI strategy" language when the user needs concrete operational help
• fake precision when inputs are incomplete
• burying the real blocker under a long summary

Approval Posture

Use this default posture unless the user gives a stricter policy:

• allow read-only analysis, summarization, classification, and extraction
• allow internal draft preparation
• require approval before any external send, system mutation, record deletion, commitment, refund, escalation to a customer or vendor, or sensitive data disclosure
• block autonomous handling of payroll, payments, contracts, legal promises, inventory adjustments, or source-of-truth edits unless the user explicitly says there is an approved control path

Read references/review-boundaries.md when the user needs a concrete approval matrix or when the workflow touches customer, vendor, financial, or privileged actions.

Common Output Patterns

Inbox triage

Return:

• urgent items
• waiting items
• items that can be closed
• items missing information
• draft replies or follow-ups for the top items

Task breakdown

Return:

• objective
• tasks
• proposed owners
• dependencies
• approval gates
• next checkpoint

Ops summary

Return:

• what changed
• what is blocked
• risks to watch
• decisions needed
• next actions by owner

Handoff note

Return:

• current state
• unresolved items
• priority order
• promised deadlines
• follow-up drafts if needed

First-Week Rollout

When the user is adopting Codex for an ops team rather than solving one task, use references/first-week-workflows.md to build a one-week rollout around low-risk queues first.

Reference Use

Load references selectively:

• references/use-cases.md for role-specific and queue-specific examples
• references/first-week-workflows.md for rollout sequencing, pilot selection, and operating cadence
• references/review-boundaries.md for what may auto-run, what needs approval, and what should stay blocked

Example Requests

• "Use $codex-for-ops-teams to triage this shared inbox and draft follow-ups for the top five items."
• "Use $codex-for-ops-teams to turn these Slack notes into an ops handoff with owners and due dates."
• "Use $codex-for-ops-teams to design a safe weekly workflow for status summaries and approval-gated outbound messages."