원클릭으로
Conjecture
Conjecture에는 kommundsen에서 수집한 skills 12개가 있으며, 저장소 수준 직업 범위와 사이트 내 skill 상세 페이지를 제공합니다.
이 저장소의 skills
Run the full release preparation checklist for Conjecture: promote PublicAPI files, update CHANGELOG, run tests, run benchmarks, update api-baseline DLLs, then output the commit message and tag commands. Use whenever the user wants to cut a new release. Triggers on "make-release", "prepare release", "cut a release", or "/make-release <version>".
Create or run a BenchmarkDotNet benchmark for Conjecture components. Use this skill whenever the user wants to measure performance, check throughput, compare two implementations, scaffold a benchmark class, or investigate memory allocations — even if they don't say "BenchmarkDotNet" explicitly. Triggers on phrases like "benchmark X", "how fast is", "measure performance of", "run benchmarks", "compare baseline vs candidate", or "check allocations".
Port a Python Hypothesis module to idiomatic C# for the Conjecture project. Use this skill whenever the user wants to translate, adapt, or port code from the Python Hypothesis library into C# — even if they don't say "port" explicitly. Triggers on phrases like "port X from hypothesis", "translate this Python module", "implement X based on hypothesis", "bring over the hypothesis shrinking logic", or when referencing a specific Python Hypothesis file or module to add to Conjecture.
Scaffold a new .NET project or module with correct structure for the Conjecture solution. Use this skill whenever the user wants to add a new project, test project, module, or set up the solution from scratch — even if they don't say "scaffold" explicitly. Triggers on phrases like "add a new project", "create a module for", "set up a test project for", "add X to the solution", or any request to create new .NET structural components in the repo.
Articulate, interrogate, and record an idea on a GitHub Discussion thread before it becomes a concrete enhancement issue. Two modes: with a discussion ref, interrogate the thread and propose follow-up comments; without a ref, collaborate on a first draft, post it to the Ideas category, then offer to enter interrogation mode. Use this skill whenever the user wants to brainstorm, develop, or pressure-test an idea, or says things like "I have an idea about X", "let's brainstorm Y", "ideate on Z", "interrogate discussion #N", "develop the idea in #N", or "I want to think through X".
Drive an entire GitHub enhancement issue to completion as one feature branch and one PR, looping through its sub-issues as TDD cycles (each committed individually). Use this skill when the user wants to ship a coherent multi-sub-issue enhancement as a single reviewable PR. Triggers on phrases like "implement enhancement #N", "work through all of #N", "drive #N to done on one branch", "one PR for the whole enhancement", or "implement #N end-to-end".
Promote a fully-interrogated and researched GitHub Discussion thread into a parent enhancement issue plus linked sub-issues, ready for `implement`. Synthesizes the discussion's idea, decisions, and research findings into the parent body, then reuses steps 7–9 of `plan-issue` (sub-issue design, creation, tasklist) — skipping the decision/question interrogation since `ideation` and `research-idea` already resolved them. Final stage of the chain `ideation` → `research-idea` → `plan-idea`. Use this skill whenever the user wants to graduate a Discussion into actual work, create the enhancement issue from a discussion, or says things like "promote discussion #N to an issue", "plan idea #N", "turn #N into issues", or "/plan-idea #N".
Research a GitHub Discussion-resident idea against the codebase and prior art, then post a structured "Research findings" comment back to the thread. Spawns Explore agents for codebase reconnaissance and WebSearch for external prior art (Hypothesis, FsCheck, QuickCheck, papers, etc.). Middle stage of the chain `ideation` → `research-idea` → `plan-idea`. Use this skill whenever the user wants to validate the feasibility of an idea, gather prior art, or check what already exists in the codebase before committing to a plan. Triggers on phrases like "research discussion #N", "is idea #N feasible", "look up prior art for #N", "what's already in the codebase for #N", or "/research-idea #N".
Plan a GitHub issue for the Conjecture project: resolve design decisions and open questions interactively with recommendations, post the answers as issue comments, then break the work into linked sub-issues ready for `implement-cycle` to iterate through. Use this skill whenever the user wants to plan or break down a GitHub issue, resolve design decisions, create sub-issues, set up an implementation roadmap, or says things like "let's start on issue #N", "plan issue #N", "break #N into cycles", "what do we need to decide for #N", or "set up the next enhancement #N".
Write documentation for the Conjecture .NET library following the Diataxis framework (https://diataxis.fr/). Use this skill whenever the user wants to write, update, or reorganize docs — tutorials, how-to guides, reference material, or explanation/background pages — for the Conjecture project. Triggers on: "write a tutorial for X", "document X", "write a how-to for X", "explain why X works this way", "reorganize the docs", "add docs for X", "reference docs for X", or any request to produce DocFX markdown for the docs/site/ tree. Also use proactively when a new feature is implemented and docs haven't been written.
Suggest a descriptive git commit message for the current changes. Use this skill whenever the user asks for a commit message, wants to know what to write for a commit, or says something like "what should my commit say", "suggest a message", or "write a commit message". DO NOT actually commit — just output the message. Invoke this skill proactively whenever the user is about to commit or asks about committing.
Record an Architecture Decision Record (ADR) in `docs/decisions/` for the Conjecture project. Use this skill whenever the user wants to document a design choice, record why something was built a certain way, capture a tradeoff, or note a decision that constrains future work — even if they don't say "ADR" explicitly. Triggers on phrases like "record this decision", "document why we chose", "log the architecture decision", "we decided to use X", or when the implement skill identifies a design choice that should be captured first.