ops-oncall-log — Claude Code skill: on-call log from alerts
Pulls alerts from your monitoring platform and writes a structured on-call log to the team wiki. Useful at the end of a shift or during a review.
npx skills add c0x12c/ai-toolkit --skill ops-oncall-log -a claude-codeWhat ops-oncall-log does
The skill walks through a fixed sequence. First it checks that two MCP servers are connected, then asks about log frequency (daily or weekly), the rotation start day, which environments to include, and where the page should live. It recommends Datadog MCP for monitoring; Confluence (mcp__claude_ai_Atlassian__*) and Notion (mcp__claude_ai_Notion__*) are the supported wiki options. If either server is absent, the skill stops with an error and writes nothing.
Alert events get grouped by base monitor name: how many times each fired, the timestamps, and the current status. Recovery events are dropped, so only triggered and warn entries count. Pages follow the hierarchy On-call Logs/On-call <YYYY>/On-call <MM>/<YYYY>/[On-call] <Start Date> - <End Date>, and the alert table carries Cause and Resolution columns. Cause is deliberately left blank for the engineer to fill in after the fact. Monitors still firing get “Still active” in Resolution.
When a page for the period already exists, the skill asks whether to update it or start a new one. On update it uses the last timestamp on the page as the starting point, so alerts already logged are not fetched again. Settings from earlier runs are reused, for instance from .spartan/config.yaml, and you confirm them before anything is pulled. The date range is always confirmed with you first.
What it helps with
- build an on-call log from monitoring alerts
- hand off a shift with a written incident history
- prepare a summary for the weekly on-call review
- spot which monitors fire most often
- flag active alerts the next engineer should pick up
Example prompts for the agent
- Build the on-call log for this week from Datadog alerts and put it in Confluence
- Update the log page for the current period with alerts since the last update
- Create a daily log for prod only and flag monitors still in Alert status
How to install ops-oncall-log
1Skills CLI (Vercel)
npx skills add c0x12c/ai-toolkit --skill ops-oncall-log -a claude-code2Manually: put the files in place
~/.claude/skills/ops-oncall-log/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/ops-oncall-log/
1Skills CLI (Vercel)
npx skills add c0x12c/ai-toolkit --skill ops-oncall-log -a codex2Manually: put the files in place
.agents/skills/ops-oncall-log/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add c0x12c/ai-toolkit --skill ops-oncall-log -a cursor2Manually: put the files in place
.cursor/skills/ops-oncall-log/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install c0x12c/ai-toolkit ops-oncall-log2Manually: put the files in place
.github/skills/ops-oncall-log/SKILL.md1Skills CLI (Vercel)
npx skills add c0x12c/ai-toolkit --skill ops-oncall-log -a gemini-cli2Platform CLI
gemini skills install https://github.com/c0x12c/ai-toolkit --path toolkit/skills/ops-oncall-logManually: put the files in place
.devin/skills/ops-oncall-log/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add c0x12c/ai-toolkit --skill ops-oncall-log -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about ops-oncall-log
How do I install ops-oncall-log in Claude Code?
Run npx skills add c0x12c/ai-toolkit --skill ops-oncall-log -a claude-code. The skill lands in ~/.claude/skills/ops-oncall-log/SKILL.md.
What does ops-oncall-log do?
It fetches alert events from your monitoring platform, groups them by monitor, and creates an on-call log page on the team wiki with an alert table and a list of active monitors.
What do I need before running it?
Two MCP servers: one for monitoring (Datadog recommended) and one for the wiki, either Confluence or Notion. Without both, the skill stops with an error.
