Install targets and dependencies

Where each install target puts skills and subagents, the by-hand steps for Claude Code, Cursor, Kiro and agents without a skills folder, what --local does, and each skill's dependency install.

The paths and per-tool details behind Install a skill into your IDE. Commands and flags are on Install options.

Manual install

Each skill is a plain directory, and installing one is always two steps: copy the skill folder into your IDE's skills or rules location, then install the skill's dependencies (the packages are in manifest.json under deps; see Installing dependencies). You can also copy the companions listed under related in the skill's manifest.json. Skills reference each other, and a hand-off to an uninstalled sibling degrades to inline guidance, but the pipelines work best complete.

Skill targets

Pick one target per install, update or uninstall:

Claude Code

Claude Code reads skills from two locations:

• User scope (available in every project): ~/.claude/skills/<skill-name>/
• Project scope (tracked with the repo): <project>/.claude/skills/<skill-name>/

Copy the skill directory itself into the skills location, not its parent pack folder:

# user scope (for personal use)
mkdir -p ~/.claude/skills
cp -R packs/solution-architect/skills/architecture-diagrams ~/.claude/skills/
cp -R packs/solution-architect/skills/figma-diagrams ~/.claude/skills/

# project scope (for sharing with a team)
mkdir -p .claude/skills
cp -R packs/solution-architect/skills/architecture-diagrams .claude/skills/
cp -R packs/solution-architect/skills/figma-diagrams .claude/skills/

Claude Code discovers the skill through the name field in its SKILL.md frontmatter. Invoke it in chat with /<skill-name> or by describing the task, and Claude routes to the matching skill.

Cursor

Cursor has no native skills concept; a skill installs as a project rule. --ide cursor does these steps for you. By hand:

1. Copy the skill folder somewhere in the repo, for example .cursor/skills/<skill-name>/:

mkdir -p .cursor/skills
cp -R packs/solution-architect/skills/architecture-diagrams .cursor/skills/

2. Create .cursor/rules/<skill-name>.mdc that points Cursor at it:

---
description: <paste the skill's description from manifest.json>
globs:
alwaysApply: false
---
Follow the instructions in .cursor/skills/<skill-name>/SKILL.md when the user requests this task.

3. In chat, attach SKILL.md with @ or describe the task; the rule fires when the description matches.

Kiro (IDE and CLI)

Kiro has native Agent Skills, and Kiro IDE and Kiro CLI read the same directories. Copying the folder in is the whole install:

mkdir -p .kiro/skills                  # workspace scope
cp -R packs/solution-architect/skills/figma-diagrams .kiro/skills/

mkdir -p ~/.kiro/skills                # global scope, every project
cp -R packs/solution-architect/skills/figma-diagrams ~/.kiro/skills/

Kiro matches the skill by its SKILL.md frontmatter name (which must equal the folder name) and description, the same contract every other tool uses.

No steering file needed. Earlier versions of the CLI also wrote .kiro/steering/<skill-name>.md pointing back at the skill. That predates native Agent Skills, and because a steering file without frontmatter is always loaded, it pinned one description per installed skill into every session's context, to point at a folder Kiro already reads. The CLI no longer writes them, and install/uninstall remove any it wrote before. A steering file it didn't author is left alone, with a note.

Codex, Antigravity and GitHub Copilot

Each has its own flag. Without --project the skills go to your personal folder and work in every repo; with it, they go into the repo:

npx skilldrop-cli install --pack dev-team --ide codex              # ~/.codex/skills
npx skilldrop-cli install --pack dev-team --ide antigravity        # ~/.gemini/antigravity-cli/skills
npx skilldrop-cli install --pack dev-team --ide copilot            # ~/.copilot/skills
npx skilldrop-cli install --pack dev-team --ide copilot --project  # .github/skills
npx skilldrop-cli install --pack dev-team --ide codex --project    # .agents/skills

All three also read other tools' folders, so a project install often reaches more than one:

If you already ran skilldrop install --project, Copilot CLI can use every skill you installed, because .claude/skills/ is one of its discovery paths. Codex and Copilot also read a repo-root AGENTS.md. Google retired Gemini CLI for most users on 2026-06-18; Antigravity CLI replaced it, and --ide antigravity is its target.

Local installs in a repo you don't own

--local installs into the repo (the same place --project would) and lists every file it writes in the repo's own .git/info/exclude. Git ignores those files the way it ignores .gitignore entries, but the exclude file is never committed, so git status stays clean and nothing can be committed by accident:

npx skilldrop-cli install --pack dev-team --local
npx skilldrop-cli uninstall pre-merge-review --local   # removes the files and the exclude entries

It works with every --ide. With --with-hooks on Claude Code, the session-start hook goes into .claude/settings.local.json rather than the shared settings.json.

Continue, Cline, Aider, and other agents

These tools don't have a standard skills directory yet. Two patterns work:

• Context attachment. Copy the skill folder anywhere in the repo, then attach SKILL.md to your prompt (Continue: @file, Cline: @file, Aider: /add <path>) and tell the agent to follow it.
• Custom prompt or agent. Paste SKILL.md into the IDE's custom-agent or system-prompt configuration. The description field in the skill's manifest.json is a good seed for the agent's name and summary.

In every case the scripts run from the copied folder, so keep the directory structure intact: don't flatten scripts/ or templates/ out of the skill folder.

VS Code (Continue and Cline extensions)

These behave like the other-agents path above. For Continue, you can also add the skill folder to .continue/config.json under contextProviders so SKILL.md shows up in @ suggestions.

Subagent targets

--agent installs reviewer subagents. Six targets, each projecting only as much as the tool's format demands:

The Kiro emitter maps Read/Grep/Glob/Bash to read/grep/glob/shell against Kiro's built-in tool reference, and names any tool it can't map instead of dropping it silently: a mistranslated permission is worse than a missing one.

Each generated target omits the permission field it cannot map safely rather than guessing one: Kiro's allowedTools and Codex's sandbox_mode are both left unset, so an agent inherits the session's permissions and is prompted per call. A widened permission nobody asked for is worse than an extra prompt.

Every surveyed tool now installs. Only Cursor is absent, because it has no agent file format at all; use a custom mode (agents/README.md has the steps).

Installing dependencies

Each skill declares its deps in manifest.json:

• deps.npm: run npm install <packages> before using the skill (or let SKILL.md step 1 install them on demand).
• deps.pip: run python3 -m pip install -r <skill>/requirements.txt.

Per-skill quick reference:

For figma-diagrams, you also need a Figma Personal Access Token exported as the FIGMA_TOKEN env var. What each skill's scripts need beyond packages is on Skills that ship scripts.

Edit this page on GitHub · View source