SuperAPI AI Directory

aisdlc-project-discover-ops-evidence

General & other

View repository

Updated 10-16 09:56

About this skill

aisdlc-project-discover-ops-evidence (Step 6: Ops & Evidence entry points and evidence chain)

Overview

The high ROI of Ops lies in: pinning down the "operational entry points" (dashboard/alerts/runbook/rollback entry points) and completing the evidence chain, reducing search costs during incidents.

Writing constraints: short, executable, provide only entry points and key points, do not copy operational manuals.

At the start, announce:"I am using the aisdlc-project-discover-ops-evidence skill to complete the ops entry points and connect the evidence chain."

Output files (optional but strongly recommended)

  • .aisdlc/project/ops/ entry pages under the directory (you can minimize according to the team's existing system):
    • ops/index.md: main entry point (Use index.md, do not use README.md)
    • Others: entry pages for monitoring/alerts/rollback/on-call/Runbook (as needed)

Minimal template (copyable)

ops/index.md

  • Operational entry point list (each item must be locatable):
    • Dashboard:<link 或配置位置>
    • Alerts:<link 或配置位置>
    • Logs:<查询入口或配置位置>
    • Runbook:<入口页或外部链接>
    • Rollback:<入口页或策略链接>
  • "First hop" for common failures:
    • Entry point (where to click)
    • Key metrics (what to look at)
    • Trigger action (what to do next/who to contact)

Evidence chain (must be reflected on the entry page)

Link the following chain together with links on appropriate pages (not required to be fully written on one page, but must be traceable):

  • Contracts → contracts/*
  • Code (implementation entry point) → code directory/route/handler/job
  • Tests (verification entry point) → test directory/command
  • CI (gate entry point) → pipeline configuration/key jobs
  • Ops (operational entry point) → dashboard/alerts/runbook/rollback

Common mistakes

  • Writing ops as a complete operational manual: project-level ops should only provide entry points and key points; details will get out of control and expire.
  • Treating migration steps as a project-level long-term asset: migration details are often one-off and prone to expiring; at the project level, keep only entry points and strategy.

Red flag checklist (if any one appears: stop and correct)

  • Under .aisdlc/project/ops/ , adding runbooks/、migrations/ and other directories and migrating the full details into them
  • The ops page has no clickable/locatable entry points at all, leaving only empty phrases like "check logs/contact ops"
  • Broken evidence chain (cannot find entry links for contracts/implementation/tests/CI)