vmihalis

Evaluator Agent — Development agent for Claude Code

Development community

Tests one attack surface for vulnerabilities — spawned per-surface with injected context from the orchestrator.

How to install Evaluator Agent

Installs to ~/.claude/agents/vmihalis-hacker-bob-evaluator-agent.md

Terminal

mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/vmihalis/hacker-bob/HEAD/.claude/agents/evaluator-agent.md -o ~/.claude/agents/vmihalis-hacker-bob-evaluator-agent.md

Restart Claude Code, or start a new session, for it to be picked up.

What Evaluator Agent does

name: evaluator-agent description: Tests one attack surface for vulnerabilities — spawned per-surface with injected context from the orchestrator tools: Bash, Read, Grep, Glob, mcp__hacker-bob__bob_ingest_sarif, mcp__hacker-bob__bob_read_static_analysis_index, mcp__hacker-bob__bob_record_candidate_claim, mcp__hacker-bob__bob_list_candidate_claims, mcp__hacker-bob__bob_repo_docker_run, mcp__hacker-bob__bob_repo_check, mcp__hacker-bob__bob_read_session_nucleus, mcp__hacker-bob__bob_write_wave_

Alternatives in Development

Recon Ranker — Attack surface ranking agent 4.5k ★
Review Workflow Context Budget — Reviews workflow machinery for context-window budget on three axes: always-loaded surface, load-on-demand disc 432 ★
.github/workflows/redteam.yml — name: Continuous Red Team Assessment on: push: branches: [main, develop] pull_request: branches: [main] schedu 213 ★