Skills DirectorySkills Directory
SkillsLearnSecurityCategoriesDocsCommunityBlog
Sign InSubmit Skill
Skills Directory

Security-tested agent skills for Claude, coding agents, and AI workflows.

Directory

  • Browse Skills
  • All Skills A–Z
  • Claude Skills
  • Claude Code Skills
  • Agent Skills
  • Categories
  • Submit a Skill

Learn

  • Learn Hub
  • Install Claude Skills
  • Write SKILL.md
  • Skills vs MCP
  • Directories Compared

Security

  • Security
  • Methodology
  • Secure Claude Skills
  • Security Badges

Company

  • About
  • Community
  • Blog
  • API Docs
  • Advertise

2026 Skills Directory. All rights reserved.

Back to skills

Stack Detection

ASecurity

Establish what a frontend project actually uses — framework, router, package manager, styling, component layer, TypeScript posture — from package.json, lockfiles, config files and the source itself, before loading framework-specific guidance or writing a line of code. Fires at the start of any frontend task in an unfamiliar or half-remembered repo, and whenever you are about to assume a convention. Not for choosing a stack for a new project, and not a substitute for reading the code you are a...

46 stars
0 votes
0 copies
0 views
Added 9/22/2026
ai-agentstypescriptgonodefrontenddocumentation

Works with

cli

Security Analysis

A100/100

Scanned 9/22/2026

Install to Claude Code

$npx -y skills add nahid-sparktales/agent-dispatcher --skill stack-detection --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Stack Detection?

Add the live security badge to your README — it updates automatically with every re-scan.

Security grade badge for Stack Detection
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/nahid-sparktales-stack-detection/badge)](https://www.skillsdirectory.com/skills/nahid-sparktales-stack-detection)

More formats (shields.io, HTML) on the badges page.

Download Zip
Files
SKILL.md
---
name: stack-detection
description: Establish what a frontend project actually uses — framework, router, package manager, styling, component layer, TypeScript posture — from package.json, lockfiles, config files and the source itself, before loading framework-specific guidance or writing a line of code. Fires at the start of any frontend task in an unfamiliar or half-remembered repo, and whenever you are about to assume a convention. Not for choosing a stack for a new project, and not a substitute for reading the code you are about to change.
---

# Stack detection

Most wrong frontend edits are correct code for a stack the project does not use. Detection is
cheap; guessing is not.

## When this fires

Before the first edit in a frontend repo you have not already mapped in this session, and again
whenever you catch yourself about to say "this project probably uses". It does not fire for a repo
whose conventions you have already read this session.

## Procedure

1. **Find the right `package.json` first.** In a monorepo the root one describes the workspace,
   not the app. Locate the nearest `package.json` above the file you are changing, and note
   whether a root one also exists — dependencies can be hoisted, tooling config usually is.
2. **Read dependencies and scripts together.** `dependencies` and `devDependencies` say what is
   installed; `scripts` say what is actually run. The dev, build and test commands are the ones to
   use later — do not invent `npm run dev` if the project spells it differently.
3. **Get real versions from the lockfile, not the range.** `^15.0.0` in `package.json` is a
   constraint. The lockfile (`package-lock.json`, `pnpm-lock.yaml`, `yarn.lock`, `bun.lock*`) holds
   the resolved version, and its filename is also the only reliable statement of which package
   manager to invoke. Two lockfiles present is a finding, not a choice.
4. **Read the config files that exist, and note the ones that do not.** Build/framework config,
   TypeScript config, lint and format config, test config. Their presence tells you the tool is
   wired in; their contents tell you path aliases, strictness, and what is excluded.
5. **Confirm the routing and rendering model from the directory layout**, not from the framework
   name. Which route directories exist, where the entry file is, whether routes are file-based or
   declared in code, and — for frameworks with a server/client split — which files carry the
   directives that mark the boundary.
6. **Identify the styling layer and its version.** A config file's existence does not pin a major
   version; check the lockfile and how the stylesheet imports the framework. Also check for a CSS
   modules / styled-in-JS / vanilla CSS layer coexisting with it — mixed styling is common and
   changes what you should write.
7. **Check for a vendored component layer.** `components.json` or an equivalent marker plus a UI
   directory means components were copied in and are now project source (see `shadcn-ui`).
8. **Open two or three real source files** in the area you will touch. Config says what is
   possible; the source says what this team actually does — import style, file naming, test
   colocation, state library usage.
9. **Write the stack down in one short block, with the file each claim came from, and a separate
   list of what you could not determine.** That list is part of the result.

## What a signal does and does not license

- A framework in `dependencies` → the framework is installed. It does **not** tell you which
  router, rendering mode or directory convention is in use. Read the layout.
- A config file exists → that tool runs. It does **not** tell you the major version, and majors
  move config between files. The lockfile decides.
- TypeScript installed → files are typed. It does **not** mean `strict` is on, that `any` is
  discouraged, or that generated types are current. Read `tsconfig` and one real file.
- A test framework installed → tests can run. It does **not** mean they pass, cover this area, or
  are run in CI. Running them is a separate act.
- A state or data library installed → it is available somewhere. It does **not** mean the module
  you are editing uses it. Grep the directory you are in.
- `components.json` present → a component CLI was used at some point. It does **not** mean the
  components are unmodified, current, or the only UI layer.
- A lockfile entry → what would install. `node_modules` is what actually runs; if behaviour
  contradicts the lockfile, the installed tree is the tiebreaker.

## Checklist

- [ ] The `package.json` read is the one governing the file being changed
- [ ] Package manager identified from the lockfile, not from habit
- [ ] Framework and styling **versions** taken from the lockfile
- [ ] Routing/rendering model confirmed from the directory layout
- [ ] Dev, build and test commands quoted from `scripts`
- [ ] At least two real source files read for convention
- [ ] Unknowns written down rather than filled in by assumption

## Failure handling

- **No lockfile** — versions are unknown. Say so; do not resolve the range in your head. Behaviour
  that depends on a major version has to be checked against the installed tree or the docs.
- **Conflicting signals** (two lockfiles, a config for a tool that is not installed, a framework
  version that contradicts the layout) — report the conflict. Picking one silently makes you the
  source of the next bug.
- **Vendored or patched dependencies** (`patches/`, resolutions/overrides, a checked-in fork) —
  upstream documentation may not describe what runs here. Read the patch.
- **Detection disagrees with the user** — say what you found and where, then follow the user.
  Do not quietly rewrite their description of their own project.

## Evidence to report

A short stack block — package manager, framework and version, router/rendering model, styling
layer and version, component layer, TypeScript posture, test setup, dev/build/test commands — with
the file each line came from, followed by an explicit "not determined" list. A stack summary with
no file references is a guess with formatting.

Attribution

nahid-sparktalesnahid-sparktales
View sourceMore from nahid-sparktales →
SSkills DirectorySkills Directory

Ship a skill? Prove it's safe.

Free 120-pattern security scan, letter grade, and an embeddable README badge.

Submit a skill

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments (0)

No comments yet. Be the first to comment!

SSkills DirectorySkills Directory

Ship a skill? Prove it's safe.

Free 120-pattern security scan, letter grade, and an embeddable README badge.

Submit a skill

Related Skills

Caveman

Ultra-compressed communication mode that cuts output tokens while keeping technical accuracy. Levels: lite, full, ultra and the wenyan variants. Use for /caveman, "caveman mode", "talk like caveman", "be brief" or "less tokens".

1066601 votes

Hyperplan

Adversarial multi-agent planning skill. Self-orchestrates 5 hostile category members (unspecified-low, unspecified-high, deep, ultrabrain, artistry) via team-mode for ruthless cross-critique debate, distills only the defensible insights, then MANDATORILY hands the distilled insight bundle to the `plan` agent for executable plan formalization. Use when planning needs maximum rigor and surfacing of weak assumptions, blind spots, and over-engineering. Triggers: 'hyperplan', 'hpp', '/hyperplan', ...

686011 votes

Mcp Code Execution

Routes multi-tool workflows through MCP servers for large datasets and pipelines. Use when Bash tool overhead is limiting throughput on data-heavy tasks.

3351 votes

catchup

Recovers the conversation and failed tool calls of a previous Codex, Claude Code, Antigravity, Cline, Copilot CLI, Cursor, DeepSeek Harness, Kimi, OpenCode, Pi Agent, or ZCode session. Use when the user says "catch up", "what did the last session do", "get me up to speed", "I switched agents", asks to recover/summarize a previous session before continuing, or asks to diagnose or report a catchup failure. Do NOT use for the current conversation, git history, or any non-agent log.

651 votes

math-skill

A comprehensive mathematical reasoning skill for AI assistants — handles arithmetic to research-level problems with rigorous step-by-step reasoning, systematic verification, and transparent uncertainty handling

381 votes
View all in ai-agents →