kkkirito-123/zlagent/using-skills
Claude Code and agent skill by kkkirito-123/zlagent
- community
Checking that the SKILL.md file is still in the repository.
Meta skill that guides the agent on when to pick an existing skill, when to create a new one, and when to install an MCP server. Adapted from obra/superpowers' "using-superpowers" and addyosmani's "meta" skill for personal-AI use.
Install
npx vskill@latest install kkkirito-123/zlagent/using-skillsSKILL.md
using-skills
Trigger
Pick this skill when the user asks what you can do, which skill applies to their task, or when the right move is unclear. Also activate internally before starting any non-trivial task so the agent does a one-pass "do I already have a skill for this?" check.
Principle
Use the right skill for the task; install a new capability onlywhen none of the installed options fit.
Core capabilities are always available but hidden from the user view. User-installed extensions surface through /list.
Decision Tree
- Does an installed skill match the intent?
- Yes → route the turn through that skill. Done.
- No → step 2.
- Can the task be done with existing tools (
read_url,
web_search, code_execution, read_file, etc.) without a new skill?
- Yes → run it directly, no new skill needed.
- No → step 3.
- Is the workflow durable (will it repeat across sessions / days)?
- Yes → propose to author a new skill via the
skill-authoring
skill. Confirm with the user before creating.
- No → one-off execution with existing tools.
- Does the task need an external service the agent cannot reach
with current tools (e.g. GitHub API, Notion, a video download service)?
- Yes → route through the
mcp-discoveryskill: search the MCP
ecosystem, propose ONE candidate, install after confirmation.
- No → the task is probably out of scope; say so plainly.
Inputs
user_intent: a short summary of what the user is asking for.current_inventory: the harness inventory (what skills / MCP /
plugins are installed). Fetch via /api/harness/extensions or by calling tool_search when browsing built-in tools.
Steps
- Summarize user intent in one line.
- Walk the decision tree above.
- If creating or installing, confirm with the user first — never
install without a yes/no.
- Record the decision briefly if the path is non-obvious ("chose
skill X because Y").
Verification
- The routed / suggested skill actually matches the intent.
- No new skill is created when an existing one would fit.
- No MCP is installed without user confirmation.
- For skill creation, run
skill-authoringrather than hand-writing
frontmatter.
Failure Signals
- The same one-off task spawns a new skill (skill inflation).
- A skill matches by name but not by workflow, causing wrong results.
- The user says "you already have that" — signals a missed inventory
check.
- Multiple MCP installs proposed in one turn — too aggressive.
Output Format
When the user asks "what can you do", reply with a compact 3-5 line summary grounded in the live inventory, not a hardcoded list. Mention core capabilities exist without enumerating them:
Security
- Security scanPassed, 90 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingworkspace/skills/meta/using-skills/SKILL.md
Popularity
- 7-day trend score
- 0
Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/kkkirito-123/zlagent/using-skills)