Agent skill

Development

by samchon in samchon/nestia

Defines nestia implementation rules, testing standards, validation, consequence analysis, and change integrity.

MITAuto-check: notesAI & LLM Engineering

Install Development

skills CLI
$ npx skills add samchon/nestia --skill development -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install samchon/nestia development --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/samchon/nestia.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.agents/skills/development .claude/skills/development && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
development
GitHub stars
2.2k
Token cost
~5.4k tokens
SKILL.md length
2,755 words
Files
1
Skills in repo
10
Repo updated
First seen
Licence
MIT

At a glance

Defines nestia implementation rules, testing standards, validation, consequence analysis, and change integrity.

  • Works in 5 steps: Finish the complete report before… → Inspect each selected declaration and… → Write @evidence contracts/.md# in native… → …
  • AI & LLM Engineering work in your project
  • SKILL.md covers Contents, Forbidden, Work Rules and Consequence Analysis, plus 5 more sections
  • Calls pnpm, npm and go

What it does

Development is an agent skill from samchon/nestia. Defines nestia implementation rules, testing standards, validation, consequence analysis, and change integrity. Use before writing or modifying source, tests, workflows, package wiring, fixtures, generated baselines, or algorithms.

Its SKILL.md is about 5.4k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.

It sits in AI & LLM Engineering. The repository describes itself as: NestJS Helper + AI Chatbot Development. The licence is MIT.

When your agent uses it

  • AI & LLM Engineering work in your project

Example prompts

  • “Use the development skill to define nestia implementation rules, testing standards, validation, consequence analysis, and change integrity”
  • “/development”

Workflow steps

5 steps, taken from the first numbered list in SKILL.md.

  1. Finish the complete report before repairing obligations. Group missing answers, code and documentation defects, selection mistakes, and…
  2. Inspect each selected declaration and its private helpers against every applicable chapter. Fix verified defects before writing the…
  3. Write @evidence contracts/.md# in native documentation, separated from descriptive prose by a blank comment line. Address the actual…
  4. Inspect public addresses with pnpm exec evidence list --config and resolution with pnpm exec evidence inspect '' --config . Account for…
  5. Recheck the complete population after each coherent repair. Verify composed claims together, including local re-export resolution.

What it can do on your machine

Read from SKILL.md and the folder at commit d3627e8. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • pnpm
    • npm
    • go

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md. Its commands use pnpm and npm, which can reach the network depending on how they are called.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Development loads about 5.4k tokens when it runs. Until then it costs about 61 tokens; SKILL.md has 2,755 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~61
When it runs · the whole SKILL.md, loaded when a task matches
~5.4k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check: notes

The automated check noted patterns worth knowing about, such as sudo or a known installer.

  • NoteMentions a .env fileSKILL.md:39
    Keep local outputs local. Do not commit `.env`, the tarballs under `deploy/tarballs/`, or any tree the harnesses regener

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from samchon/nestia at commit d3627e8, republished under its MIT licence (© samchon). 2,755 words, ~5,399 tokens.

Download SKILL.mdSave it as .claude/skills/development/SKILL.md (or your agent's skills folder).
name
development
description
Defines nestia implementation rules, testing standards, validation, consequence analysis, and change integrity. Use before writing or modifying source, tests, workflows, package wiring, fixtures, generated baselines, or algorithms.

Development

Contents

Forbidden

Read the contracts skill before changing maintained production declarations. Its common checklist owns implementation acknowledgments; scoped checklists apply by responsibility. When a failure disproves an assumption, correct its owner and remove superseded compensations in the same repair.

These four are never acceptable; choosing any one means the approach is already wrong.

  • No monkey-patching or hardcoding. Don't special-case a consumer, a controller name, a DTO name, a fixture name, or an expected value to make output match. Fix the general logic.
  • No test-passing-only logic. Code exists to be correct, not to turn a check green. A branch whose only purpose is to satisfy one assertion is a bug in disguise.
  • No forcing a broken design. When the same failure keeps returning under patch after patch, the design is wrong. Stop, find the root cause, and fix the design instead of looping forever on symptoms.
  • No whack-a-mole. Don't patch the one case that surfaced and move on. Think expansively about every case the same root cause can produce, and seal them all with coverage so the class of failure cannot recur.

Work Rules

  • Choose the principled course. Time, difficulty, and the breadth of consequences require more careful analysis and validation; they never justify a shortcut, leaving a verified consequence unaddressed, or a weaker acceptance standard.
  • Match existing conventions. Before adding a file, function, or test, open a nearby peer in the same package and mirror its naming, location, and code style; don't create parallel structures.
  • Respect package boundaries. The Go binary lives in @nestia/core and is shared with @nestia/sdk as a linked contributor. Don't fork a second native binary, don't give @nestia/sdk its own ttsc plugin entry, and don't reintroduce a TypeScript-side transformer.
  • Keep detection general and installation-safe. The native transform must locate target packages through their resolved package root, as packages/core/native/transform.cjs does with require.resolve("@nestia/sdk/package.json", { paths }) and packages/sdk/src/transform.ts with createRequire(...).resolve("@nestia/sdk/package.json"). Workspace-only path substrings break the moment a consumer installs from npm. The same rule applies to the generators: no hard-coded DTO, controller, or feature names.
  • Preserve the public contract in .agents/skills/project/SKILL.md. Decorator names, INestiaConfig options, CLI flags, and the generated SDK / Swagger / e2e surface are public; renaming or removing any of them is a deliberate product change, not incidental cleanup.
  • Use the workspace catalogs. pnpm-workspace.yaml pins versions under catalog:typescript, catalog:samchon, catalog:nestjs, catalog:utils, catalog:modelcontextprotocol, and catalog:rolldown. New dependencies go through the matching catalog; internal references use workspace:^.
  • Migration templates ship without interactive dependencies — generated projects stay non-interactive.
  • Keep local outputs local. Do not commit .env, the tarballs under deploy/tarballs/, or any tree the harnesses regenerate (tests/test-e2e/.tmp-*, generated consumers and benchmark reports under the integrated E2E workspace).
  • When public behavior changes, update the matching page under website/src/content/docs/** in the same change. Follow .agents/skills/documentation/SKILL.md.
  • Run pnpm format once on the final pull-request changes before merge and include its result in that pull request. Formatting is not a prerequisite for each commit or push, including campaign commits. If later edits change a formatter target, rerun it before merge; an unchanged final snapshot needs no repeat. One invocation covers both the package and test trees, so no workspace needs formatting separately. All other validation and merge gates remain required.

Consequence Analysis

Treat a reported example as one witness of a cause, not the complete problem statement. Before changing code, trace the same cause through:

  • every caller and downstream consumer, including the generated SDK, Swagger document, mockup simulator, and e2e suite;
  • normal, error, and recovery state transitions;
  • both HTTP adapters — a decorator or transform change that behaves differently on Express and Fastify is two behaviors, not one;
  • concurrency, caching, and generated output;
  • Windows and POSIX behavior;
  • compatibility constraints and boundary inputs.

Fix the verified class of failure, not only the reported witness. Cover positive, negative, and boundary cases without expanding the user's product goal.

Plugin Configuration

ttsc discovers nestia from the ttsc.plugin.transform entry in packages/core/package.json, which points at packages/core/native/transform.cjs. That descriptor owns the composition model documented in .agents/skills/project/SKILL.md: the cmd/ttsc-nestia Go source, composes: ["typia/lib/transform"], and the conditional @nestia/sdk contributor.

Keep type analysis, AST rewriting, diagnostics, and transform options on the Go side. Consumer projects add a compilerOptions.plugins entry only when they need to set optional transform flags.

Test workspaces point at @nestia/core/native/transform.cjs. Do not add a second plugin host, or a separate @nestia/sdk plugin entry, to make a local layout pass.

Testing

One test case per file, named after what it asserts. This is the rule for new and changed cases in both languages, even where older files predate it. Nothing enforces it mechanically — DynamicExecutor gates on the test prefix only and will happily run a file exporting three functions or one whose export name disagrees with its filename — so the discipline is yours to keep.

Go tests

Go unit tests live in two dedicated package modules, not beside the source:

  • packages/core/test, run by pnpm --filter @nestia/core test:go.
  • packages/sdk/test, run by pnpm --filter @nestia/sdk test:go.

Each module carries replace directives back to ../native (and, for the SDK, to ../../core/native) plus the pinned typescript-go shim redirects. Use one Test* function per file, named after the assertion, and mirror a nearby test's package, fixture, and cleanup pattern. Exercise option, analysis and emit semantics through the owning operations in-process; reserve native-binary or installed-CLI execution for the necessary wrapper connection in the E2E population. Do not rebuild or launch a host per rule.

Generated-validator runtime connections share the SDK integration preparation. Pure option selection, diagnostics and declaration provenance belong to the core unit module; a native dispatch plus Node execution is an integration boundary even if its test file is Go. Count those actual operations separately from test-language preparation.

Native production trees contain no *_test.go; root pnpm test:go runs the two owning test/ modules. Keep new tests there unless same-package access is necessary and wire any exceptional test location into the canonical runner. Package compilation remains a prerequisite, not an additional empty test population.

Every test:go script passes -count=1, and so should a hand-run go test. Cacheable mode instruments the test binary to record every file a test opens, and these tests open the whole TypeScript program and its node_modules typings, which on Windows made one package take minutes instead of seconds. A cached pass would also be stale, because the tests read the pnpm-installed toolchain and fixtures outside Go's view.

TypeScript suites

The TypeScript workspaces are test-benchmark, test-cli, test-e2e, test-editor, test-migrate and test-sdk. tests/test-e2e contains only direct units of packages/e2e; do not place SDK, migration or other integration cases there. SDK and migration workspaces separate their direct unit entries from necessary integration connections. test-transform-options is retired: native option semantics belong to the owning Go units and necessary installed-loader/runtime connections share the SDK integration preparation. Empty legacy entries must not produce a vacuous pass.

Classify by the real call path. Consumer installation, separate product compilation, Nest application creation, HTTP hosts, worker sessions and CLI/IPC processes are integration preparation. Running TypeScript test source is language preparation; directly calling a parser, composer or writer with authored input remains unit logic when it does not create those integration boundaries. Loading an already-built internal operation by absolute path does not itself make the direct unit E2E.

Function-per-file unit entries use DynamicExecutor under src/features/**; editor browser cases use src/browser/features/** in a separate process to prevent DOM-dependent initialization leaking into SSR. Export exactly one test_<snake_case> function from a matching filename. Name helpers outside the discovered prefix, derive the source extension from __filename and reject zero discoveries.

The sole E2E entry owns shared preparation and teardown. Combine compatible connections into one rich authored input program and one generated consumer, using public ttsc/ttsx or TtscCompiler. Do not preserve per-feature compiler contexts behind two Node processes, create arbitrary shards, invoke the old workspace starts or introduce a dedicated SDK Go host. Transfer exact pure name/schema/option judgments to their owning units when independent-program premises conflict with a shared fixture.

Record every original valid assertion's owning operation, independent oracle, normal/error controls, unit or integration destination and preparation costs before replacing its execution path. Rich migration inputs must retain request/response, security, multipart, plain-text and schema distinctions; lowering fixture thresholds or treating compile success as those assertions is invalid. Failed preparation and tests retain their first result, and cleanup must cover partial startup.

Keep code under a JSDoc @example unfenced. The JSDoc formatter can emit a closing fence with a trailing semicolon and move acknowledgment tags behind it, making the checker read those tags as example code. Use a description-level fenced block when a displayed fence is needed, and separate acknowledgment tags from prose with a blank comment line.

Use the shared helpers in @nestia/e2e and each suite's local internal/ helpers. Do not reach into another suite's internals.

Open every new or materially changed case with a doc comment in the same three-part shape: a one-line Verifies ... headline, a short paragraph stating the non-obvious why (which branch or regression is being pinned), and a 2+-step numbered list summarizing the scenario.

ts
/**
 * Verifies @TypedBody with `validate: "assertPrune"` strips extras from both
 * input and return value.
 *
 * Locks the assertPrune branch of the native body validator. Other validate
 * modes pass the input through untouched; only the Prune family removes
 * extras, and the transform has to pick the matching typia helper. A
 * regression in helper selection would silently fall back to a non-pruning
 * validator and let extra properties through.
 *
 * 1. Build a controller method with `@TypedBody()` and `validate: "assertPrune"`.
 * 2. Send a payload carrying an extra property.
 * 3. Assert the extra is missing from both the input and the return value.
 */
export const test_body_config_assertPrune_strips_extras = (): void => {
  /* ... */
};
Coverage, not happy paths

A test that only feeds a controller its ordinary valid input and asserts a 200 proves one path, not correctness. The transform spans validators, serializers, Swagger metadata, SDK emit, mockup simulators, and diagnostics; each predicate or branch needs more than its happy path:

  • The transformation direction. When a call should be rewritten, assert observable emitted or runtime behavior that differs from the untransformed stub. For the generators, start from a hand-written controller and verify the generated SDK, Swagger, or e2e output — not merely that an existing output still compiles.
  • A negative twin for every positive. Wherever a predicate accepts, rewrites, narrows, serializes, or reports a diagnostic, pin an adjacent case one property away where it must not do so. An over-match stays invisible until the counter-example exists. Rejected authored inputs in the owning unit or shared E2E fixture supply the same negative controls.
  • Boundaries. Cover the empty case, the single-element case, recursion limits, optional and nullable members, exact numeric limits, deepest nesting, and the decorator option or compiler flag that flips the decision.
  • Oracle-derived expectations. Take expected behavior from TypeScript semantics, the NestJS contract, and the authoritative OpenAPI specification — never from whatever the current code happens to emit. A snapshot written against the implementation's own output locks its bugs in.
Show full SKILL.md (1,050 more words)Show less

Validation

Run the narrowest command that proves the change first, then a broader command when shared behavior, the transform, packaging, or documentation changed. Report any command that could not be run.

  • One Go module: pnpm --filter @nestia/core test:go or pnpm --filter @nestia/sdk test:go; root pnpm test:go runs both modules.
  • One TypeScript workspace: pnpm --filter ./tests/<name> start.
  • Unit population: pnpm test:unit runs the populated pure TypeScript unit workspaces against caller-built artifacts. Editor browser initialization stays in its additional process. Continue independent populations after a failure; no unit entry installs consumers, compiles product fixtures or starts hosts, workers or CLI/IPC sessions. pnpm test:go runs the native unit modules through owning operations in-process.
  • E2E population: after pnpm build, run pnpm test:e2e, whose SDK module entry selects the SDK and migration owners independently. tests/test-e2e remains exclusively the direct packages/e2e unit population. Count actual installation, compiler, generation, backend and worker operations within each integration owner. test.yml owns all tests in one job after one installation and package build, with distinct Evidence, Go, JavaScript-unit and integration steps and no matrix or shards; measure its complete duration against the eight-minute target while preserving assertions.
  • One package: pnpm --filter ./packages/<name> build.
  • Transform, decorators, or generators broadly: pnpm test; use pnpm build as the faster compilation gate.
  • Packaging: run root pnpm package:tgz, then inspect or smoke-test a clean install.
  • Website or guide changes: run root pnpm install and pnpm run package:tgz first, because website/package.json depends on ../deploy/tarballs/editor.tgz and ../deploy/tarballs/migrate.tgz. Then run npm install --force && npm run build inside website/, which also rebuilds @nestia/migrate and @nestia/editor and runs TypeDoc into public/api.

Verification shape depends on the change type:

  • Bug fix: name the failing case and the expected behavior; run a repro that fails before the fix and passes after.
  • Feature: name the observable behavior; exercise it end-to-end.
  • Refactor: name what should stay unchanged; rely on the existing test suite or a behavior-locking probe.
  • Review: name concrete risks, missing tests, or regressions.

Each shared E2E producer and consumer phase runs once with visible diagnostics. A failed compilation or intermittent transport result remains a failed feature while unrelated features continue; do not repeat project preparation merely to reveal its output or replace an earlier failure with a later pass. The eight-minute CI target requires shared preparation and preserved first-failure evidence.

Change Integrity

Treat tests, fixtures, snapshots, CI workflows, package wiring, dependencies, core algorithms, generated baselines, and benchmark results as part of the specification. Changing them requires an explicit user request or a clear product reason, and the final report must call it out.

Go source under packages/*/native must ship under a version newer than the latest published packages. The npm packages ship their native/ trees and ttsc compiles that source on first use, so without a new package version npm consumers keep installing the previous native source tree.

One maintainer-owned release change assigns that version, where bumpp -r moves all eight packages together. Multiple unreleased native changes belong to the same future release instead of consuming one patch number each, and no implementation pull request changes a version field. A release change may begin only after campaign completion, or after the user explicitly suspends the campaign and lifts the freeze, and it uses the version the user assigned.

For mechanical ports, migrations, or broad rewrites, preserve the existing algorithm and public behavior in reviewable slices. Prefer a concrete exemplar over abstract instructions, and inspect the diff before trusting a green test run.

Evidence Adoption

Each package's evidence.config.json selects the files that directly declare its maintained TypeScript types and functions and, for @nestia/core and @nestia/sdk, the Go types and functions under native/, and references contracts/common.md under ../../.agents/skills. Pure re-export barrels own no selected declaration and are excluded; review their public wiring through package and consumer checks while selecting each symbol at its direct declaration. A file that mixes a declaration with a foreign re-export or an ambient declare module block is split so the declaration owner stays selected, because the checker resolves only relative modules inside the selected source root. Hand-written .d.ts files for untyped dependencies own no implementation and are excluded. Properties keep native documentation and are reviewed through their type. Add the portability and performance chapters only to the operations that own those decisions. Each tests/test-* workspace owns its own evidence.config.json and evidence script, with module-relative source selections and contract references under ../../.agents/skills. Tests always answer contracts/testing.md; an actual E2E boundary also answers contracts/e2e.md.

Run root pnpm evidence to collect shared-config, published-package and test-workspace owners without stopping at the first failure. The ordinary recursive pnpm command runs each module's own Evidence script and configuration and aggregates their statuses. Use pnpm --filter <package> evidence for one module and pnpm evidence:tests for all test workspaces. JSON configuration keeps the checker independent of the native plugin this repository ships. .github/workflows/test.yml runs the command in its own step and has no path filter, so every pull request checks the full contract surface. Acknowledgments do not replace behavioral tests.

The root deploy/** scripts are top-level script bodies with no exported declaration the adapter can address, so they are review-only.

  1. Finish the complete report before repairing obligations. Group missing answers, code and documentation defects, selection mistakes, and incomplete analysis by cause.
  2. Inspect each selected declaration and its private helpers against every applicable chapter. Fix verified defects before writing the answer; the acknowledgment describes the resulting implementation.
  3. Write @evidence contracts/<document>.md#<anchor> <reason> in native documentation, separated from descriptive prose by a blank comment line. Address the actual question for that declaration. Use @evidenceExclude only for a genuinely inapplicable individual chapter with its concrete reason.
  4. Inspect public addresses with pnpm exec evidence list --config <config> and resolution with pnpm exec evidence inspect '<target>' --config <config>. Account for private helpers, anonymous callbacks, dynamically registered cases, script entry bodies, and compile-only cases that the adapter cannot address. Make executable entries selectable where practical; record remaining review-only coverage honestly.
  5. Recheck the complete population after each coherent repair. Verify composed claims together, including local re-export resolution.

Do not weaken severity, narrow a maintained population, add generic compliance prose, or exclude a whole document to silence obligations. Exclude generated output, copied fixture input, dependencies, and build output only with verified provenance. Authored generators and helpers remain maintained source. Verify selection against the actual runners rather than inferring enrollment from globs.

© samchon, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

Just SKILL.md in .agents/skills/development of samchon/nestia.

Open the folder on GitHubat commit d3627e8

Compare with similar skills

Development next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Development compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Development this skillsamchon/nestia2.2k—~5.4kAutomated safety check: NotesMIT
Doc AuthorInsForge/InsForge13k—~3.7kAutomated safety check: PassMIT
Add AI Chat Toolryokun6/ryos1.3k—~2.2kAutomated safety check: PassAGPL-3.0
Benchmarksamchon/typia5.9k—~1.2kAutomated safety check: PassMIT
Agent Interface DesignNeeeophytee/finding-unknowns-skills343—~650Automated safety check: PassMIT
Search Benchmarkagentic-community/mcp-gateway-registry962—~1.9kAutomated safety check: PassApache-2.0

Similar skills

  • Doc Author

    InsForge/InsForge

    Write, edit, and maintain documentation. An agent skill from InsForge/InsForge.

    13k GitHub stars~3.7k tokensUpdated yesterday
    AI & LLM EngineeringAuto-check passed
  • Add AI Chat Tool

    ryokun6/ryos

    Add or modify an AI chat tool ("Ask Ryo" capability) in ryOS.

    1.3k GitHub stars~2.2k tokensUpdated today
    AI & LLM EngineeringAuto-check passed
  • Benchmark

    samchon/typia

    Defines typia benchmark fixture integrity, result reporting, and publication safeguards.

    5.9k GitHub stars~1.2k tokensUpdated yesterday
    AI & LLM EngineeringAuto-check passed
  • Agent Interface Design

    Neeeophytee/finding-unknowns-skills

    Design tools, scripts, and CLIs that an agent will call, so the interface teaches its own use instead of a wall of prose and examples.

    343 GitHub stars~650 tokensUpdated 9 days ago
    AI & LLM EngineeringAuto-check passed
  • Search Benchmark

    agentic-community/mcp-gateway-registry

    Generate a search quality benchmark for the AI Registry. An agent skill from agentic-community/mcp-gateway-registry.

    962 GitHub stars~1.9k tokensUpdated yesterday
    AI & LLM EngineeringAuto-check passed
  • Implement Feature

    ljxpython/ai-agent-platform

    AI 在项目文档(docs/projects/{YYYYMMDD}-{项目名}/)已存在时,实现任务过程中自动调用(不需要用户手动触发),记录改动细节到 implementation/ 目录。跳过条件:单项目改动的小改动不需要创建实现记录。

    137 GitHub stars~1k tokensUpdated yesterday
    AI & LLM EngineeringAuto-check passed

More from samchon/nestia

All 10 skills in this repo
  • Project

    samchon/nestia

    Defines the nestia product contract, workspace layout, package boundaries, the Go plugin composition model, and canonical commands.

    2.2k GitHub stars~3k tokensUpdated today
    Auto-check passed
  • Contracts

    samchon/nestia

    Defines self-acknowledgments for production declarations and tests.

    2.2k GitHub stars~1.3k tokensUpdated today
    Auto-check passed
  • Issue Campaign

    samchon/nestia

    Defines the default solo repository-wide issue campaign for nestia: exhaustive discovery, lead-vetted issue publication, one unified CI-validated implementation pull request per cycle, solo…

    2.2k GitHub stars~3k tokensUpdated today
    Auto-check passed
  • Pull Request

    samchon/nestia

    Defines nestia branch, commit, pull-request, check, and merge workflows.

    2.2k GitHub stars~1.6k tokensUpdated today
    Auto-check passed
  • Review

    samchon/nestia

    Defines exhaustive solo review, Self-Review, and solo repository-wide issue-discovery rounds for nestia.

    2.2k GitHub stars~2.1k tokensUpdated today
    Auto-check passed
  • Discussion

    samchon/nestia

    Runs structured multi-agent discussions for open-ended nestia topics.

    2.2k GitHub stars~424 tokensUpdated today
    Auto-check passed

Questions about Development

What does Development do?

Defines nestia implementation rules, testing standards, validation, consequence analysis, and change integrity. Development is an agent skill from samchon/nestia. Defines nestia implementation rules, testing standards, validation, consequence analysis, and change integrity.

When should I use Development?

Development fits situations like: AI & LLM Engineering work in your project.

How do I install Development in Claude Code?

Run `npx skills add samchon/nestia --skill development -a claude-code`. Or copy the skill folder (.agents/skills/development in samchon/nestia) into .claude/skills/development in your project. Claude Code loads it when a task matches its description.

How do I install Development in Codex?

Run `npx skills add samchon/nestia --skill development -a codex`. Or copy the skill folder (.agents/skills/development in samchon/nestia) into .agents/skills/development in your project. Codex loads it when a task matches its description.

Can I use Development in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add samchon/nestia --skill development -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/development, .gemini/skills/development, .github/skills/development and .opencode/skills/development in your project.

What does Development need to run?

Going by SKILL.md and its folder, Development needs the command-line tools its instructions call (pnpm, npm and go).

Does Development access the network?

SKILL.md contains no URLs. Its commands use npm, which can reach the network depending on how they are called. This is read from the text; nothing was executed.

Is Development safe to install?

Our automated static check of SKILL.md found notes only (mentions a .env file), nothing it rates as a warning. It is not a guarantee. Review the folder before installing.

What licence does Development use?

Development is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Development use?

About 5.4k tokens (SKILL.md is roughly 22k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Development?

Skills that share tags, products or a category with Development: Doc Author (InsForge/InsForge, 13k stars), Add AI Chat Tool (ryokun6/ryos, 1.3k stars), Benchmark (samchon/typia, 5.9k stars) and Agent Interface Design (Neeeophytee/finding-unknowns-skills, 343 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Development?

samchon (a GitHub user) maintains it in samchon/nestia, which has 2,179 GitHub stars. The repository holds 10 skills in this directory. The repository was last updated on October 7, 2026.

Source: samchon/nestia on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.