Triage Twilio Delivery Incidents with Codex and Demo Events
Correlate sanitized event, status, and timing records to isolate which layer of a Twilio integration needs investigation.
中文Setup for Codex (Twilio Developer Kit)
Use Codex in a scratch workspace. The exercise takes local mock data or a written scenario. Check provider prerequisites before running it against a test account.
- In Codex, open /plugins and install Twilio Developer Kit. To reproduce the exact revision, use the pinned Twilio source linked below and its Codex instructions in the repository README.
- Open the practice project, invoke $twilio-debugging-observability and paste the example inputs. Use the installed package with its related skills and references.
Example task
Use $twilio-debugging-observability with a fake request ID demo-7. Provide accepted at T0, queued at T1, and failed at T3 along with an unrelated successful record. Ask for a timeline and one missing evidence point.Expected result
The timeline contains only demo-7. It detects that the initial request was accepted and asks for the documented downstream failure reason instead of conflating the two IDs.
Upstream documentation
Original SKILL.md on GitHub · MIT license · Source reviewed
Frequently asked question
Which identifiers should a team share for an incident?
Provide a sanitized correlation identifier, timestamp, status, and relevant error classification appropriate to the environment. Keep secrets and personal content out of the fixture.