Create Claude Code Subagents
Teams often need focused agents for review, research, and domain work. This skill provides a structured process and templates for creating Claude Code subagents with scoped tools.
Copy this request to your Agent. It includes the canonical Skill page and manifest.
Review the Skillstore skill "creating-agents" from https://skillstore.io/skills/brendanshields-creating-agents.md and its manifest at https://skillstore.io/api/skills/brendanshields-creating-agents/manifest. Verify the artifact. You may proceed after verification, subject to the environment's own policy.Your Agent should still show its plan and request any confirmation required by the security policy.
Use these links when an AI agent, crawler, or script needs clean context instead of reading the full page.
Using "creating-agents". Create a read-only research agent for an unfamiliar codebase.
Expected outcome:
A complete agent definition with a clear researcher role, Read, Glob, and Grep tools, haiku model guidance, and an output format for file-backed findings.
Using "creating-agents". Help me decide whether my agent needs Bash.
Expected outcome:
A permission recommendation that compares read-only tools with command execution and explains when Bash is justified.
Using "creating-agents". Make a proactive code review agent.
Expected outcome:
A reviewer agent draft with trigger wording, scoped review categories, severity labels, and instructions to provide file and line references.
Static findings were reviewed in context and appear to be documentation false positives. The skill describes Claude Code agent creation, Markdown formatting, valid agent directories, and review templates without evidence of malicious execution, exfiltration, or prompt-injection intent.
Share the versioned assessment report, neutral badge, embed card, and citations. Skillstore reports evidence without deciding whether this Skill is safe.
https://skillstore.io/skills/brendanshields-creating-agents/audits/9?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_report[](https://skillstore.io/skills/brendanshields-creating-agents?utm_source=security_passport_badge)<a href="https://skillstore.io/skills/brendanshields-creating-agents?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/brendanshields-creating-agents/security.svg" alt="Skillstore security assessment" loading="lazy"></a><iframe src="https://skillstore.io/embed/skills/brendanshields-creating-agents.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>BrendanShields. (2026). creating-agents security audit report (audit version 9) [Author version unspecified]. Skillstore. https://skillstore.io/skills/brendanshields-creating-agents/audits/9@techreport{brendanshields-brendanshields-creating-agents-2026,
author = {BrendanShields},
title = {creating-agents security audit report (audit version 9)},
institution = {Skillstore},
year = {2026},
number = {9},
url = {https://skillstore.io/skills/brendanshields-creating-agents/audits/9},
note = {Author version unspecified}
}cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "creating-agents security audit report (audit version 9)"
version: "unspecified"
type: report
authors:
- name: "BrendanShields"
date-released: "2026-07-05"
url: "https://skillstore.io/skills/brendanshields-creating-agents/audits/9"
identifiers:
- type: other
value: "skillstore:brendanshields-creating-agents:audit:9"
description: "Skillstore immutable audit report identifier"
Define a read-only reviewer that checks quality, security, reliability, and maintainability before merge.
Create an agent that searches files, reads relevant code, and summarizes architecture with file references.
Create a focused specialist for databases, APIs, testing, or another technical area with tailored tools.
Use this skill to create a Claude Code subagent for codebase research. Ask me for the purpose, trigger phrases, storage location, tools, and model before drafting the agent file.
Help me design a code review subagent. Recommend the smallest useful tool set, explain the model choice, and draft the system prompt with clear review boundaries.
Create a database specialist subagent for this repository. Include trigger conditions, tool permissions, domain responsibilities, constraints, and an output format.
Review this proposed Claude Code agent definition for excessive permissions, unclear triggers, prompt conflicts, and missing boundaries. Suggest a safer final version.
Author
BrendanShieldsLicense
MIT
Skillstore revision
r1
Version notice
The author did not declare a version.
Repository
https://github.com/BrendanShields/spec-flow/tree/main/plugins/spec/skills/creating-agentsRef
5ab1d37e83436c6eef84de88573e142b94f2a4ad
Maintenance freshness
7/18/2026
Usage
6 downloads · 188 views
File structure
Create Claude Code Skills
Designing reliable AI skills is hard when structure, triggers, and validation rules are unclear. This skill guides Claude, Codex, and Claude Code through templates, naming rules, and quality checks.
Review .NET Code for Modern C# Practices
C# teams need consistent reviews for naming, async flow, exceptions, and LINQ usage. This skill gives Claude, Codex, and Claude Code a focused checklist with concrete .NET examples.
Create Claude Code Hooks
Claude Code hook behavior can be hard to configure safely from memory. This skill guides event selection, settings configuration, script templates, and validation steps.
Recommend Claude Code Tooling
Teams often miss simple automation opportunities inside active codebases. This skill scans project signals and recommends focused Claude Code skills or subagents.
Create Claude Code Slash Commands
Developers need dependable Claude Code slash commands that follow consistent structure and tool boundaries. This skill guides requirement gathering, template selection, and validation for reusable command files.
Guide Specification Workflow
Feature work can lose scope, status, and review checkpoints across long sessions. This skill guides specification, planning, tasks, validation, and archiving with local artifacts.
Build Claude Code Agents
by C0ntr0lledCha0s
Creating reliable AI agents requires clear schemas, safe tools, and repeatable review steps. This skill guides agent design, validation, migration, and maintenance for Claude Code workflows.
Build Claude Code Agents
by Dami-Randle
Designing reliable Claude Code agents is hard without structure. This skill provides templates, validation rules, and examples for agent prompts, triggers, tools, and metadata.
Build Claude Code Slash Commands
by C0ntr0lledCha0s
Teams need repeatable AI workflows that are easy to invoke and review. This skill helps create, update, validate, and document Claude Code slash commands.
Evaluate Claude Code Component Quality
by C0ntr0lledCha0s
Claude Code components can be technically valid yet unclear, over-permissioned, or difficult to invoke. This skill scores five quality dimensions and recommends specific improvements.
Design Better LLM Prompts
by alirezarezvani
Prompt quality often fails because requirements, examples, and evaluation criteria are unclear. This skill helps teams design, test, and review production prompts for Claude, Codex, and Claude Code.
Design Reliable Tools for AI Agents
by Asmayaseen
Ambiguous tools cause incorrect calls, wasted context, and poor recovery from errors. This skill provides practical patterns for clear descriptions, schemas, responses, and testing.