Evaluate intent and assertions in specification tests with dry-run IR debugging.
npx skills add https://github.com/ianphil/my-skills --skill spec-tests
Please help me install this Agent Skill. Skill: spec-tests Source: https://github.com/ianphil/my-skills/tree/main/spec-tests Command: npx skills add https://github.com/ianphil/my-skills --skill spec-tests
Intent-based specification tests address the gap between what code does and why it matters, ensuring LLM-driven development preserves business goals and prevents testers from altering thresholds or requirements to pass.
Create a spec file with a frontmatter target, an H3 test name, an explicit intent above its code block, and a fenced assertion block.
High-intent search queries and answers about installing and using this skill.
Intent-based specification tests evaluate both why a feature matters and what it must do, preventing testers or LLMs from altering thresholds to game the system and ensuring development preserves business goals.
Create a spec test file by defining a frontmatter target, adding an H3 test name, stating an explicit intent above the code block, and including a fenced assertion block to enforce per-test intent.
Yes, you can debug failed specification tests using the provided IR or dry-run view. The framework supports rerunning failed tests and offers this intermediate representation to inspect multi-target specs during debugging.
You prevent LLM-generated code drift by adopting intent-driven spec tests that enforce dual evaluation of both feature intent and assertions, bridging the gap between what code does and why it matters for quality assurance.
Yes, intent-testing supports multi-target specs within a single file. You can define multiple targets in the frontmatter and evaluate their specific intents and assertions, rerunning only the failed tests as needed.
easonlaoValidate `.claude/skills/*/SKILL.md` files for structural compliance and behavioral correctness.
tiweenCreate a Trigger Map aligning business goals with user psychology.
brsblGenerate per-file documentation with narrative history from commit changes.
EntityProcessCreate and manage AgentV YAML evaluation files for AI agent performance testing.
