ticket-reviewlisted
Install: claude install-skill arthaszyb/bright-talent
# Ticket Review (Cache-Cluster Scaling)
Review cache-cluster scaling change tickets raised in the **Change Gateway**
(mocked for this demo). Fetch the ticket and its 7-day peak metrics, apply the
cache-scaling SOP, and post a structured pass/warn/fail review comment.
**This skill never approves or rejects a ticket.** It produces a comment for a
human reviewer to act on.
## When to trigger
Trigger on:
- A Change Gateway ticket URL, e.g. `https://gateway.acme.example/tickets/1002`
- A bare ticket ID with review intent, e.g. "review this scaling ticket 1002"
- Pasted ticket JSON/content with a request to review it
Do **not** trigger (or trigger as review) on:
- Pure status queries ("what's the status of ticket 1001") — that's a lookup,
not a review; a review re-runs the SOP checks and produces a fresh comment
- Any request to approve, reject, or otherwise act on the ticket
("approve ticket 1001") — this skill has no approve/reject capability by
design; redirect the user to the Change Gateway's own approval action
## Inputs
- **Ticket ID** (required): extracted from a Change Gateway URL
(`https://gateway.acme.example/tickets/{id}`), a bare ID, or pasted content.
- **Base URL** (optional): defaults to `http://localhost:8801`, the bundled
mock server. Override only if the user names a different mock endpoint.
## Script pipeline
All deterministic work — fetch, parse, compute, render — is scripted. The
LLM's job is to orchestrate these four steps and present the res