Validate Claude Code skills using the Kiln task-shapes testing framework.
npx skills add https://github.com/yoshisada/ai-repo-template --skill task-shapes
Please help me install this Agent Skill. Skill: task-shapes Source: https://github.com/yoshisada/ai-repo-template/tree/main/plugin-kiln/lib/task-shapes Command: npx skills add https://github.com/yoshisada/ai-repo-template --skill task-shapes
Validates Claude Code skills through a standardized Kiln task-shapes testing framework.
Use Kiln to run the task-shapes tests via /kiln:kiln-test <plugin> <test> and review the pass/fail results.
High-intent search queries and answers about installing and using this skill.
You can validate Claude Code skills by applying the Kiln task-shapes testing framework, which defines standardized task shapes across agent, backend, cli, data, docs, frontend, infra, and skill tests to ensure consistent end-to-end evaluation scenarios.
Run Kiln tests by executing the command /kiln:kiln-test <plugin> <test> through the Kiln plugin. This verifies skill coverage and reports exit codes, the last PASS/FAIL line, and transcript-based diagnostic metrics.
Yes, the Kiln task-shapes framework covers end-to-end skill evaluation scenarios across agent, backend, cli, data, docs, frontend, infra, and skill tests, providing a consistent test harness for comprehensive coverage.
Kiln testing reports exit codes, the last PASS/FAIL line, and transcript-based metrics. These diagnostic outputs help you verify skill coverage and identify specific failures during the evaluation process.
Yes, task-shapes facilitates QA automation for Claude Code skills by defining standardized task shapes within a testing harness. It ensures consistent evaluation scenarios run via the Kiln plugin to verify skill behavior.
A standardized testing framework ensures consistent evaluation scenarios across multiple domains like agent, backend, and infra tests. Without it, validating Claude Code skills end-to-end lacks uniform metrics for pass/fail verification.
bitwize-music-studioExecute pytest suites across 14 categories to validate Claude Code plugin integrity.
bernajaberRun and validate automated acceptance tests for product features.
AradotsoCoordinate Claude Code subagents for specialized development workflows.
CDO-07-NewDefine repeatable evaluation harnesses for Claude Code task outcomes.
