Crew Agentic — Testing skill for Claude Code
An AI-powered coding agent, similar to Claude Code, that autonomously handles end-to-end software development, including project scaffolding, feature implementation, bug fixing, refactoring, testing.
How to install Crew Agentic
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open harsh409/-crew_agentic and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Crew Agentic does
An AI-powered coding agent, similar to Claude Code, that autonomously handles end-to-end software development, including project scaffolding, feature implementation, bug fixing, refactoring, testing, and project task execution.
Alternatives in Testing
- Factory — Dark Factory Mode - Spec-in, software-out autonomous pipeline 2.8k ★
- Create Test Repo — Create a new ArcKit test repo with full scaffolding, auto-detecting the next version number 2.2k ★
- Spec Kitty — Spec-Driven Development for serious software developers 1.6k ★
README
CrewAI Coding Assistant
A multi-agent coding assistant built on **CrewAI**, running against your local LLMAPI gateway. This is the CrewAI rewrite of the hand-built project
How it maps
| Piece | File | Job |
|---|---|---|
| The LLM every agent uses | llm.py |
wraps LLMAPI as a CrewAI LLM (LiteLLM, openai/ prefix) |
| Safe file layer | workspace.py |
the only thing that touches the target repo; every path is jailed |
| The agents' hands | tools.py |
CrewAI @tools: list / search / read / edit / create / run / check / google_search |
| The agents | agents.py |
Distributor (manager) + Explainer, Scope, Coder, Tester, Reviewer |
| The crew | crew.py |
one hierarchical crew: the Distributor delegates to the right specialists |
| Sessions | sessions.py |
save / list / continue past chats |
| Token saver | token_saver.py |
condenses long history so big chats stay cheap |
| CLI entry point | main.py |
tasks, questions, sessions, interactive |
Long / multi-line prompts
The terminal's one-line input can't handle a big prompt (blank lines end it, paste breaks). Use any of these instead:
# 1. from a file (best for long prompts) - write it once, reuse it
python3 main.py -p ~/Desktop/app -f prompt.txt
# 2. from stdin
python3 main.py -p ~/Desktop/app - < prompt.txt
cat prompt.txt | python3 main.py -p ~/Desktop/app
# 3. inside interactive mode, open your $EDITOR to compose it
python3 main.py -i -p ~/Desktop/app
Task> /edit # opens nano/vim; write freely, save, close
Task> /file prompt.txt # or load a prompt file
Modes and options
| Command | What it does |
|---|---|
python3 main.py -p "" |
run an edit-task (Scope → Coder → Tester → Reviewer) |
python3 main.py -p "" |
a question is auto-detected and answered read-only |
python3 main.py -p --question "" |
force answer-only mode (never edits) |
python3 main.py -i -p |
interactive chat; every turn |
Related Skills
Buildwright
Approve intent, ship autonomously. Agent-first autonomous development workflow for Claude Code, OpenCode, and
Test Driven Development.SKILL
Enforces TDD discipline with RED-GREEN-REFACTOR cycle. Use when writing new features, fixing bugs, or refactor
Accenture Enterprise Multi Agent Code Review Orchestrator
A production-ready multi-agent code review system powered by the Claude Agent SDK. Specialized AI agents colla
Fix Bug
Test-driven bug fixing — reproduce with a failing test, fix via parallel agents, then pick the best solution.
Al Debug
Root-cause a bug/test failure/race condition before fixing; checks and maintains a project-local known-issues
Claude Engineer Toolkit
AI-powered toolkit for Claude Code that provides structured workflows, commands, and agents to automate planni
Related Agents
Deepseek Coder
DeepSeek coding agent for implementation, code modification, refactoring, test execution and fixing implementa
Senior Developer
Use this agent when the user needs programming tasks executed, including feature implementation, refactoring,
Loom Software Engineer
Primary implementation agent for all coding work - features, bug fixes, tests, data pipelines, ML training, in