Resource Manager — Testing & QA agent for Claude Code
Use BEFORE dispatching any TEST-RUNNING subagent (a builder, a fix-subagent, qa-engineer) to get a GO/HOLD verdict with the number of free slots against MEASURED host capacity; AFTER agents finish to.
How to install Resource Manager
Installs to ~/.claude/agents/vlad-ko-claude-wizard-resource-manager.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/vlad-ko/claude-wizard/HEAD/agents/resource-manager.md -o ~/.claude/agents/vlad-ko-claude-wizard-resource-manager.md Restart Claude Code, or start a new session, for it to be picked up.
What Resource Manager does
name: resource-manager description: Use BEFORE dispatching any TEST-RUNNING subagent (a builder, a fix-subagent, qa-engineer) to get a GO/HOLD verdict with the number of free slots against MEASURED host capacity; AFTER agents finish to REPORT which per-worktree test resources have become reclaimable; and on ANY suspicion of a hung agent, an OOM, a full data volume, lock contention, or a cluster of failures that might be infrastructure rather than code. It also answers what nothing else in th
Alternatives in Testing & QA
- Staff Code Reviewer — Use for adversarial pre-merge code review 179 ★
- Filament — Filament v4 admin panel specialist 154 ★
- Eval Improver — Improver role of the D365FO agent eval loop 136 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Joe Pipeline
Orchestrates the full review -> fix -> test -> publish cycle by dispatching
Cha Mutant Smith
Writes mutants that break one meaning-carrying line each, with anchors lifted verbatim and a real test path. U
LLM MCP
Engineer of redbar's AI surface — the MCP server, the handoff that hands the gap to the agent to write the tes
Task Fixer
Repair agent that reads spec.md, verdict.json, and problems.md, applies the smallest safe fix, and refreshes e
Fix Applier
Use this agent after root-cause-analyst has issued a verdict with a spec-only or guarded fix scope, to make th
QA Judge
Independent read-only evaluator. Scores another agent's output (scenarios, healing fix, triage verdict) agains
Related Skills
Airport Investment Agent
AI agent that helps an investor pick which US airports to invest in. Public BTS/FAA data pipeline, determinist
Hub Hold
Hold a shared resource for this chat across several steps (Unreal Agent Hub)
Se Agent Control Hold Finish
agent control / hold: sets stop.hold to finish. What the session does when it reaches the end of a turn.
