From agentic
Design and implement Claude Code plugin skills end-to-end: analyze existing plugin conventions, spec invocation contract and workflow, confirm design with the user, then write SKILL.md plus any required reference/context files and validate consistency. Use for new skills, major rewrites, or expanding a plugin's capability surface.
How this skill is triggered — by the user, by Claude, or both
Slash command
/agentic:build_skill [plugin-name-or-path] [skill-name-or-description][plugin-name-or-path] [skill-name-or-description]The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Spec and build a new Claude Code plugin skill. `$ARGUMENTS` should include the target
Spec and build a new Claude Code plugin skill. $ARGUMENTS should include the target
plugin (name or path) and the skill name or short description. If either is missing,
ask the user before continuing.
Fetch the following before doing any analysis:
Read these before continuing. Do not rely on training-time knowledge when the docs are available.
Before speccing anything, study the existing plugin anatomy:
Resolve the plugin path using this order:
.claude/plugins/<name>/ in the current project<name>/.claude-plugin/ relative to cwd~/.claude/plugins/cache/ latest matching entryRead these files if present:
.claude-plugin/plugin.jsonskills/*/SKILL.mdskills/*/reference.mdcontext/*.mdreferences/*.mdhooks/hooks.jsonCapture conventions from the plugin before proceeding:
name, description, argument-hint)If conventions are inconsistent, note that and propose one consistent style in the spec.
Work through this checklist with the user:
$ARGUMENTS be interpreted?After gathering answers, provide a complete spec summary and ask the user to confirm.
Present this template and ask for explicit confirmation before writing files:
Skill name: <agentic:...>
Directory: skills/<name>/
Main file: skills/<name>/SKILL.md
Supporting files: <none | list>
Purpose:
<one sentence>
Invocation contract:
$ARGUMENTS -> <parsed form>
Workflow:
1. <step>
2. <step>
Outputs:
- <deliverable>
Dependencies:
- <files/skills>
Edge cases:
- <case> -> <behavior>
Ask: "Does this spec look right before I write the files?"
Write in this order:
skills/<name>/SKILL.mdUse valid frontmatter and a clear step-based workflow.
Required frontmatter keys:
namedescriptionRecommended:
argument-hintBody structure:
Only create extra files if the spec requires them:
skills/<name>/reference.md for large static guidancereferences/<file>.md for reusable plugin-wide docscontext/<file>.md for session context templatesDo not add files that are not justified by the confirmed spec.
If the plugin has a documented index/list of skills, update it. If there is no index convention, do not invent one.
After writing files, run lightweight checks:
skills/<name>/SKILL.md.name in frontmatter matches intended invocation.Report any mismatches and fix them before finishing.
skills/<name>/SKILL.md (required)npx claudepluginhub nominexhq/agentic-plugin --plugin agenticProvides behavioral guidelines to reduce common LLM coding mistakes, focusing on simplicity, surgical changes, assumption surfacing, and verifiable success criteria.
Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.
Creates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.