AgentStack
SKILL unreviewed MIT Self-run

Agent Skills Setup

skill-luckycat133-skills-repo-skills-repo · by Luckycat133

>

No reviews yet
0 installs
0 views
view→install

Install

$ agentstack add skill-luckycat133-skills-repo-skills-repo

Open-source listing — not yet scanned by AgentStack. Follow the source repository for install instructions.

Security review

⚠ Flagged

1 finding(s); flagged for manual review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures
  • high Pipes remote content directly into a shell (remote code execution).

What it can access

  • Network access Used
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

View the full security report →

Reliability & compatibility

Not yet reviewed
0 installs to date
no reviews yet
yesterday

Declared compatibility

Claude CodeClaude Desktop

Compatibility is declared by the source manifest. End-to-end runtime verification is coming — see below.

Preview Execution monitoring

We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps — measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.

How agent discovery & health will work →
Are you the author of Agent Skills Setup? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

AI IDE Context Migration

Migrate AI assistant context (MCP, rules, skills, commands, agents, hooks, memory) between IDEs with format conversion, safe merging, and verification.

> MCP Protocol: 2025-11-25 (stable). SSE deprecated; use Streamable HTTP for remote servers. > Full IDE Registry: Read references/ide-registry.md for detailed per-IDE paths of all migration objects.


Migration Objects

| # | Object | Description | |---|--------|-------------| | 1 | mcp | MCP server configurations (stdio/HTTP) | | 2 | rules | Instructions/rules/context files | | 3 | skills | SKILL.md skill directories | | 4 | commands | Slash commands / prompt templates | | 5 | agents | Subagent definitions | | 6 | hooks | Lifecycle event hooks | | 7 | memory | Persistent memory / memory banks / context files |

Never migrate: API keys, tokens, OAuth tokens, chat history/transcripts, IDE UI settings, built-in vector indexes, workspace storage, SQLite databases.


Execution Workflow

1. DETECT    — Scan filesystem for installed IDEs using detect_dir in IDE Registry
2. IDENTIFY  — Ask user: source IDE(s) and target IDE(s)
3. SCAN      — Read references/ide-registry.md; scan ALL migration objects for source IDE
4. DRY-RUN   — Generate migration preview: list objects, conversion plan, conflicts
5. CONFIRM   — Show preview to user; wait for explicit approval before writing
6. BACKUP    — Create .bak.TIMESTAMP copies of existing target files that will be modified
7. MIGRATE   — Execute migrations with format conversion per Object Conversion Rules
8. VERIFY    — Validate output files parse correctly; run verification commands; report results

Critical rules:

  • Default to DRY-RUN. Never write without user confirmation.
  • Always backup before overwriting. Use .bak. suffix.
  • Merge, never overwrite. Conflicts renamed to _migrated.
  • Blank all secret values in env (keep key names, set values to "").
  • If source or target config is invalid JSON/TOML/YAML, STOP and report.

IDE Quick Reference

Root key and format differences — the most common migration errors:

| IDE | MCP Root Key | Format | Config Path | Key Pitfall | |-----|-------------|--------|-------------|-------------| | Claude Desktop | mcpServers | JSON | claude_desktop_config.json | MCP only; no rules/skills | | Claude Code | mcpServers | JSON | ~/.claude.json / .mcp.json | MCP NOT in settings.json | | Cursor | mcpServers | JSON | ~/.cursor/mcp.json / .cursor/mcp.json | Skills since v0.44; hooks supported | | Cline | mcpServers | JSON | globalStorage path (VS Code) | CLI uses ~/.cline/mcp.json separately | | VS Code Copilot | servers | JSON | .vscode/mcp.json | servers NOT mcpServers; needs type | | Copilot CLI | mcpServers | JSON | ~/.copilot/mcp-config.json | Needs type; no longer reads .vscode/mcp.json | | Windsurf | mcpServers | JSON | ~/.codeium/windsurf/mcp_config.json | Uses serverUrl not url | | Zed | context_servers | JSON | settings.json | 1.4.2+ uses AGENTS.md, removed @rule | | Trae | mcpServers | JSON | .trae/mcp.json | Project MCP needs manual toggle | | Trae Work | mcpServers | JSON | .trae/mcp.json | Shares Trae IDE config; SOLO/Code mode for agents | | Codex CLI | [mcp_servers.*] | TOML | ~/.codex/config.toml | Underscores in names; prompts deprecated→skills | | Gemini CLI | mcpServers | JSON | ~/.gemini/settings.json | Server names must use hyphens; commands are TOML | | OpenCode | mcp | JSON | opencode.json | mcp not mcpServers; command is array; env is environment | | Goose | extensions | YAML | ~/.config/goose/config.yaml | extensions not mcpServers; cmd/args/envs | | JetBrains Junie | mcpServers | JSON | ~/.junie/mcp/mcp.json | Rules is .junie/guidelines.md single file | | Kiro | mcpServers | JSON | .kiro/settings/mcp.json | Path has settings/ subdir; supports hooks | | Continue.dev | mcpServers | YAML | ~/.continue/config.yaml | mcpServers is ARRAY not object | | Sourcegraph Amp | CLI-managed | JSON | ~/.config/amp/ | MCP via amp mcp add, not config key | | Sourcegraph Cody | mcpServers | JSON | ~/.config/cody/mcp_servers.json | Free/Pro sunset 2025-06; MCP via feature flag; agentic gathering not @mentions | | PearAI | mcpServers | JSON | ~/.pearai/config.json | Standard object format; file is config.json | | Forge | mcpServers | JSON | ~/.forge/.mcp.json | Path is ~/.forge/ not ~/.config/forge/ | | Void Editor | mcpServers | JSON | ~/.config/void/mcp_servers.json | File is mcp_servers.json not mcp.json | | Kilo Code | mcpServers | JSON | ~/.config/kilo/kilo.jsonc | Global is ~/.config/kilo/; agents from modes | | Roo Code | mcpServers | JSON | .roo/mcp.json | Archived 2026-05; migrate to Kilo Code | | Amazon Q | mcpServers | JSON | ~/.aws/amazonq/default.json | Legacy mcp.json with useLegacyMcpJson flag | | antigravity | mcpServers | JSON | ~/.gemini/config/mcp_config.json | Config NOT in antigravity/ (that's cache) | | Augment Code | mcpServers | JSON | ~/.augment/settings.json | GUI primary; skills follow agentskills.io | | OpenHands | mcpServers | JSON | ~/.openhands/mcp.json | CLI 1.0+ JSON; GUI uses config.toml [mcp] | | Aider | N/A | YAML | ~/.aider.conf.yml | No native MCP; rules via CONVENTIONS.md | | Helix | [mcp_servers.*] | TOML | ~/.config/helix/config.toml | Via helix-ai plugin; tools-only | | Tabnine | mcpServers | JSON | .tabnine/mcp_servers.json | Rules via guidelines/rules.md | | Tongyi Lingma | mcpServers | JSON | .lingma/mcp-settings.json | GUI primary; ModelScope integration | | Baidu Comate | mcpServers | JSON | .comate/mcp.json | Rules use .mdr format; supports .agents/skills/ | | Tencent CodeBuddy | mcpServers | JSON | ~/.codebuddy/.mcp.json | Full rules/agents/skills/memory system | | Kimi Code | mcpServers | JSON | ~/.kimi-code/mcp.json | Path is .kimi-code NOT .kimi; config is config.toml | | ZCode (Zhipu) | mcp.servers | JSON | ~/.zcode/cli/config.json | Root key mcp.servers (dot); uses AGENTS.md not CLAUDE.md | | MiniMax Code | N/A | — | (desktop app) | Built-in Agent Team/Skills/Memory; MCP unconfirmed | | mmx-cli (MiniMax) | N/A | JSON | ~/.mmx/config.json | Region trap: api.minimax.io vs api.minimaxi.com | | Qoder CN (fmr Tongyi Lingma) | mcpServers | JSON | ~/.qoder/ / qodercli mcp add | Renamed 2026-05-20; uses AGENTS.md; ModelScope MCP plaza | | Baidu Comate IDE | mcpServers | JSON | (desktop app) | Distinct from plugin; Zulu multi-agent; .mdr rules; global config since 2025-08 | | Tencent CodeBuddy IDE | mcpServers | JSON | (desktop app) | Distinct from plugin; shares config with CodeBuddy CLI | | iFlyCode (讯飞) | mcpServers | JSON | (UI config) | MCP via UI paste JSON; no file-level config docs | | Raccoon AI (商汤) | mcpServers | JSON | (UI config) | MCP via UI paste JSON; no file-level config docs | | MonkeyCode (长亭) | mcpServers | JSON | ~/.monkeycode/ | AGPL-3.0 open source; SDD specs; MonkeyScan; private deploy | | veCLI (火山引擎) | mcpServers | JSON | ~/.vecli/ | Distinct from Trae CLI (Volcano Engine vs Trae brand) | | bolt.new / bolt.diy | servers | JSON | ~/.boltai/mcp.json | Root key servers NOT mcpServers!; supports import from Cursor/Claude | | Qodo (fmr CodiumAI) | mcpServers | JSON | IDE Tools Mgmt UI | CodiumAI ≠ Codeium; agents in .toml; CLI supports --mcp | | Devin (Cognition) | mcpServers | JSON | (cloud dashboard) | Cloud SaaS only; bidirectional MCP; also exposes mcp.devin.ai | | v0 (Vercel) | UI | — | v0.app/.../mcp-connections | Bidirectional MCP; also exposes mcp.v0.dev; whitelisted clients only | | Lovable | UI | — | (lovable.dev chat connectors) | Bidirectional MCP; server OAuth limited to 5 clients; Enterprise disabled | | gptel + mcp.el (Emacs) | elisp | — | ~/.emacs.d/init.el | Emacs ecosystem; mcp-hub-start; requires lisp | | WorkBuddy | mcpServers | JSON | ~/.workbuddy/.mcp.json | Path is ~/.workbuddy/ not ~/.workbuddy root; skills at ~/.workbuddy/skills/ |


Universal Standard Locations

Migrate to these FIRST for maximum cross-IDE compatibility:

| Path | Purpose | Loaded By | |------|---------|-----------| | AGENTS.md | Universal project instructions | Claude Code, Cursor, Copilot, Codex, Gemini, Zed 1.4.2+, JetBrains/Goose/Kiro/Aider/OpenCode/Augment/Forge/Void/OpenHands/Kimi Code/ZCode/Qoder CN/veCLI | | .agents/skills//SKILL.md | Universal skills | Cursor, Copilot, Codex, Trae, antigravity, OpenCode, Augment, Amp, Zed, Comate, Kimi Code | | .claude/skills//SKILL.md | Claude-compatible skills | Claude Code, Cursor (loads natively), OpenCode, Augment, Copilot | | .mcp.json | Universal project MCP | Claude Code, Copilot CLI | | .vscode/mcp.json | VS Code ecosystem MCP | Copilot, Cline, Continue, Cody, Kilo Code, Void |


Object Conversion Rules

MCP Server Conversion

CONVERT_MCP(source_config, source_ide, target_ide):

  1. Read source MCP config using source_ide paths and root_key
  2. For each server entry:
     a. Extract: command, args, env, url, headers
     b. BLANK all secret values in env (keep key names, set to "")
     c. Convert to target format per root_key:
        - mcpServers (JSON object): direct copy
        - servers (JSON, VS Code Copilot): rename mcpServers→servers; add type:'stdio'|'http'
        - context_servers (JSON): rename root key (Zed)
        - mcp.servers (JSON): rename root key (ZCode; also accepts mcpServers)
        - mcp (JSON): rename; add type:'local'|'remote'; command→array; env→environment (OpenCode)
        - mcpServers ARRAY: object → [{name,type,...}] (Continue.dev ONLY)
        - servers (JSON, bolt.new/bolt.diy): rename mcpServers→servers (NO type field needed)
        - mcp_servers (TOML): → [mcp_servers.] table; underscores (Codex/Helix)
        - extensions (YAML): rename; command→cmd, env→envs; add type (Goose)
        - amp: via `amp mcp add` CLI command (Sourcegraph Amp)
     d. Handle special fields:
        - Windsurf HTTP: url → serverUrl
        - antigravity HTTP: url → serverUrl
        - Copilot CLI: add type:'local'|'http'
        - Codex HTTP: uses url+bearer_token_env_var or http_headers
        - Gemini CLI: server names underscores → hyphens
        - Continue.dev: convert object to array [{name,type,...}]
        - Goose: add type:'stdio'|'sse'|'streamable_http'
  3. Merge with existing target config (backup first; duplicates → _migrated)
  4. Write; verify parses correctly

Rules/Instructions Conversion

Rules are MARKDOWN. The BODY is always reusable — only frontmatter and filename need adaptation. AGENTS.md is the universal intermediate format.

| Source | Conversion | |--------|------------| | .cursorrules / .windsurfrules / .clinerules / .voidrules | Copy body; rename to target's rules file | | .cursor/rules/*.mdc | Extract body; adapt frontmatter (description, globs, alwaysApply) | | CLAUDE.md / GEMINI.md / CODY.md / CODEBUDDY.md / HELIX.md | Copy body; rename to target's filename | | AGENTS.md | Copy directly (universal standard) | | .kiro/steering/*.md | Copy body; note conditional steering (inclusion: always/fileMatch/auto/manual) | | .augment/rules/*.md | Extract body; adapt frontmatter (always/auto/manual) | | .comate/rules/*.mdr | Extract body; .mdr is Markdown with Comate extensions | | .junie/guidelines.md | Copy body; single file not directory | | .tabnine/guidelines/rules.md | Copy body | | Any rules → Zed/ZCode/Kimi Code/Qoder CN/veCLI | Append to AGENTS.md (all use AGENTS.md as universal) | | Any rules → Aider | Use CONVENTIONS.md; add to read: in .aider.conf.yml | | Any rules → Goose | Copy to .goosehints or AGENTS.md |

Skills Conversion

CONVERT_SKILL(source_skill_dir, target_skill_dir):
  1. Keep: name, description (universally supported per agentskills.io)
  2. Remove IDE-specific frontmatter: allowed-tools (Claude), agent/fork/color (Augment)
  3. Copy entire skill directory (SKILL.md + scripts/ + references/ + assets/)
  4. Also copy to .agents/skills/ (universal) and .claude/skills/ (Claude-compatible)
  5. Verify SKILL.md has required name + description

Commands/Prompts Conversion

Commands are markdown files (filename = command name). Exceptions:

  • Gemini CLI: commands are .toml files, not markdown
  • VS Code Copilot: rename to *.prompt.md; add mode/model/tools frontmatter
  • Cody: convert to JSON format ({commands: {key: {prompt, description, mode, context}}})
  • OpenCode: support $ARGUMENTS, $1, !cmd, @filepath templates
  • Kimi Code: no standalone commands dir; use skills or plugin commands

Agents Conversion

Copy markdown body; adapt frontmatter. Supported fields by IDE:

  • Claude Code: name, description, model, tools
  • VS Code Copilot: name, description, tools, model
  • OpenCode: description, mode, model, tools, permission
  • Forge: name, description
  • Tencent CodeBuddy: name, description, tools, model
  • Kiro: name, description, prompt, mcpServers, tools, hooks (JSON format)
  • Kimi Code: YAML format (extend, name, systempromptpath, tools, exclude_tools, subagents)
  • Roo Code/Kilo Code: Modes system (.roomodes/kilo.jsonc) — per-mode tool permissions
  • Gemini CLI: via /agents command and extensions
  • ZCode: Markdown (global ~/.zcode/agents/ or project .zcode/agents/)
  • Qodo: TOML format (agents/.toml: instructions, tools, commands)

Hooks Conversion

| Source | Target | Conversion | |--------|--------|------------| | Claude Code (settings.json hooks) | Cursor (.cursor/hooks.json) | Adapt event names; Cursor supports sessionStart/End, preToolUse/postToolUse, etc. | | Claude Code hooks | VS Code Copilot (.github/hooks/*.json) | Shared format; 8 lifecycle events | | Claude Code hooks | Kiro (.kiro/hooks/*.kiro.hook) | Convert to JSON; when.type→fileEdited/preToolUse/postToolUse; then.type→runCommand/askAgent | | Claude Code hooks | OpenCode (.opencode/plugins/*.ts) | Convert to TypeScript event handlers | | Claude Code hooks | Kimi Code (config.toml [[hooks]]) | Convert to TOML array; 13 events; blocking: PreToolUse/Stop/UserPromptSubmit | | Any hooks | Others | Note as manual step |

Memory Conversion

| Source | Target | Conversion | |--------|--------|------------| | Claude Code (~/.claude/projects//memory/MEMORY.md + user*.md etc.) | Any rules | Copy MEMORY.md + relevant .md files as project context | | Cline memory-bank/*.md (6 files: projectbrief, productContext, activeContext, systemPatterns, techContext, progress) | Any rules | Copy as project context; prepend "## Project Context" | | Goose ~/.config/goose/memory/ (JSON shards) | Any rules | Extract text from JSON shards; convert to markdown | | Trae ~/.trae/memory/user_profile.md | Any rules | Copy directly | | Windsurf memories (~/.codeium/windsurf/memories/) | Any rules | Copy globalrules.md; other memories are auto-generated | | CodeBuddy CODEBUDDY.md | Any rules | Copy directly | | Gemini CLI /memory | Any rules | Use /memory show to export; copy to target | | MiniMax Code (3-tier memory) | Any rules | Export via MiniMax Code UI; convert to markdown | | Cross-IDE: mem0 MCP / OpenMemory / Pieces LTM | Same | Install as MCP server in target IDE; memory follows API key/device |

Key rule: Memory files are markdown/context. Convert to target IDE's rules or memory format. For MCP-based memory (mem0, Pieces), install the MCP server in target IDE.


Safety Boundaries

| Never do this | Why | |--------------|-----| | Copy/migrate API keys, tokens, secrets | Security; blank values, tell user to fill in | | Overwrite existing config without backup | Data loss; always .bak.TIMESTAMP first | | Overwrite existing entries | Merge only; conflicts → _migrated | | Execute package installs (npm install -

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet — be the first.

Versions

  • v0.1.0 Imported from the upstream source.