Most Cursor command packs are markdown snippets. You paste a prompt, hope the agent behaves, and when it regresses you shrug and rewrite the paragraph. That is fine for personal notes. It is a bad way to ship workflows other people install.
I wanted the opposite: slash commands that behave like a small product. Each /command is a thin entry. The real contract lives in a paired skill. Behavioral evals score PASS / PARTIAL / FAIL. Ship-gate fixtures in CI fail the build if a safety guard disappears from the skill text. No LLM judge on every PR. Structural anchors only: cheap, boring, enforceable.
A walk through /gauntlet-loop
Take /gauntlet-loop. The idea is simple and mean: stop saying "make it better." Beat a real example.
Command file. Thin YAML frontmatter plus Overview, Defaults, Steps, Anti-patterns, Examples. Step 1 always resolves the skill contract (workspace path, then user install fallback). Anti-patterns use a fixed shape: Trigger / Wrong / Correct / Reason. The Correct behavior must also exist as a positive guard in the skill.
Skill contract. Intake requires a GOAL and a REAL-WORLD EQUIVALENT, plus an inspectable reference pack (files, screenshots, clip, build, or repo path). A famous name alone is not a pack. The agent decomposes into independent parts, runs a per-part state machine (build → critique → pass | iterate | terminal), and never lets builders grade their own work. Critics use fresh context. Pass only if better than the reference; equal fails. After all parts pass, an integration critic grades the whole. Optional budget, taste-domain pride gate, and a gap ledger for resume.
Eval rubric. Cases cover missing reference, missing pack, equal-is-fail, skip-critic, fabricated blind, budget exhaustion, resume-without-retrying-stalled-gaps, and more. PARTIAL counts as fail on the ship gate.
Fixtures in CI.
eval/fixtures.yamllistsskill_requiredphrases that must appear verbatim inSKILL.md. Delete "Do not let builders evaluate their own work" andrun-eval-fixtures.py --strictturns red. That is the point: negative knowledge is locked, not tribal.
Why this matters for installers
When you add the repo as a Cursor user plugin, you get the contracts and the CI story, not just a menu of vibes. The catalog stays portable (generic examples, no employer names). Plugin install syncs with your account across desktop, web, CLI, and mobile.
Install: https://github.com/emaraschio/cursor-commands via Customize → Plugins. Not an official Cursor product. MIT.
If you maintain your own slash commands, the reusable idea is smaller than the whole catalog: pair every command with a skill, write a few ship-gate cases, and fail CI when the guard text disappears. That alone beats hoping the model remembers last month's prompt.
Top comments (2)
Ship-gate fixtures that fail the build when a safety guard disappears from the skill text is the most transferable idea here — structural anchors instead of an LLM judge on every PR keeps it cheap and deterministic. The critics-use-fresh-context rule matches what we converged on independently; self-grading is where agent pipelines quietly rot. How do you keep the anchors themselves from going stale when the skill text legitimately gets rewritten?
Appreciate that. Short version: anchors are co-owned with the skill. Rewrite SKILL.md → update fixtures.yaml in the same PR → --strict fails until they match. We anchor stable rule phrases, not decorative prose, and we delete fixture rows when a guard is intentionally retired. CI keeps them honest; a manual ship-gate walk covers the case where the words survive but the behavior does not.