Qwen Recon — Development agent for Claude Code
Read-only repository investigator for broad architecture and dependency mapping.
How to install Qwen Recon
Installs to ~/.claude/agents/b-a-m-n-claudebrigade-qwen-recon.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/B-A-M-N/ClaudeBrigade/HEAD/.claude/agents/qwen-recon.md -o ~/.claude/agents/b-a-m-n-claudebrigade-qwen-recon.md Restart Claude Code, or start a new session, for it to be picked up.
What Qwen Recon does
name: brigade-recon description: Read-only repository investigator for broad architecture and dependency mapping. tools: Read, Grep, Glob, Bash permissionMode: plan background: false
Inspect the repository at scale without modifying it.
Return exact files, symbols, control flow, tests, coupling, configuration boundaries, compatibility risks, and unresolved uncertainties.
Do not design the final architecture. Supply repository evidence to the main DeepSeek controller.
Alternatives in Development
- Architecture Verify — Architecture verification specialist 538 ★
- Shep Clean Arch Auditor — Read-only clean architecture auditor for shep 247 ★
- Architecture Guardian — Use to verify MacMagazine module boundaries and system compliance are intact after adding/moving code or editi 169 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Factory Investigator
Reads and judges the recon a triage investigation gathered — the inventory, the find-refs hits, the worker-exp
Signal Miner
Lowest-cost, read-only utility for commands expected to produce large logs or stdout. Use it for tests, benchm
Vzt Scout
Recon agent (Haiku 5.5) — codebase search, file location, dependency lookups, summaries, status checks, 'where
Review Architecture
Read-only architecture reviewer for medium and high complexity talk-first changes - duplication, god classes,
Bead Dependency Mapper
Use this agent when mapping bead dependencies — finding bottlenecks (open issues blocking the most other open
Goap Agent
Invoke for complex multi-step tasks requiring intelligent planning and multi-agent coordination. Use when task
Related Skills
Engage.Recon
Execute Phase 1 - Reconnaissance and Attack Surface Mapping
Long Context Retrieval
Read large files and broad search results progressively instead of blowing up the live context window.
V Epic
Drive an EPIC — chain several features into one autonomous, resumable, dependency-ordered build on a single br