cekura-bot-test-writer

Updates committed Cekura voice-agent test suites to match pull request diffs.

7|2|Updated Mar 6, 2026
One-click install
npx skills add https://github.com/cekura-ai/cekura-skills --skill cekura-bot-test-writer-cekura-ai
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: cekura-bot-test-writer
Source: https://github.com/cekura-ai/cekura-skills/tree/main/cekura/skills/cekura-bot-test-writer
Command: npx skills add https://github.com/cekura-ai/cekura-skills --skill cekura-bot-test-writer-cekura-ai

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? When a pull request changes a voice agent, the repository's committed Cekura test suite often falls out of sync, and manually deciding which test cases need updating is slow and error-prone. This Skill reads the PR diff against its merge base, decides per change whether existing cases already cover it, and makes the smallest edit that closes a real gap. ## Core Features & Use Cases - Diff-driven triage: Translates each changed hunk into transcript-observable behavior using a per-layer decision table covering prompts, providers, thresholds, tools, languages, and transports. - Minimal edit ladder: Chooses the cheapest edit mode (KEEP, TIGHTEN, EXTEND, REPLACE, RETIRE, ADD) so suites stay stable and case keys preserve their history. - Offline validation: Lints edited specs and runs a dry-run validation against the Cekura API without ever placing a live call. - Use Case: A pull request lowers the idle prompt timer from 10s to 6s. The Skill detects that the existing idle-recovery case's 8s silence now straddles the new timer, replaces the turn value in place, keeps the case key, and reports the decision table in the PR comment. ## Quick Start Ask the assistant to review this pull request and update the committed Cekura test suite for any voice agent changes it contains.

Frequently Asked Questions about cekura-bot-test-writer

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I update voice agent tests for a pull request?

Run the Skill on the PR: it diffs against the merge base, translates each changed hunk into transcript-observable behavior, and produces a decision table. It then applies the cheapest edit mode, usually no edit at all, and validates the spec with a dry run.

How do I keep a Cekura test suite in sync with code changes?

Use the Skill's edit ladder: KEEP when existing cases cover the change, TIGHTEN to add assertions, EXTEND to add turns to compatible cases, and ADD only for genuinely new subsystems. It never regenerates or renames cases, preserving result history.

Can this run automatically on every pull request in CI?

Yes, the references include GitHub Actions wiring with path filters, label triggers, and a guard step that fails the job if anything outside the spec files changes. The job validates only and never places live calls.

Does the test writer place live calls or create new test suites?

No. It never places live calls, only lints and dry-run validates edited specs. It also never creates a suite from scratch; if no committed suite exists, it hands off to the cekura-infra-test-suite skill instead.

What happens when a PR makes an existing test assertion fail?

The Skill distinguishes deliberate product changes from regressions. If the diff does not settle which it is, it stops and reports the case, assertion, and breaking hunk rather than weakening the assertion to make the suite pass.