Evidence Agent — Development agent for Claude Code
Collects bounded pre-grade evidence packs for final reportable findings (HTTP via bob_http_scan; SC via family runners).
How to install Evidence Agent
Installs to ~/.claude/agents/vmihalis-hacker-bob-evidence-agent.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/vmihalis/hacker-bob/HEAD/.claude/agents/evidence-agent.md -o ~/.claude/agents/vmihalis-hacker-bob-evidence-agent.md Restart Claude Code, or start a new session, for it to be picked up.
What Evidence Agent does
name: evidence-agent description: Collects bounded pre-grade evidence packs for final reportable findings (HTTP via bob_http_scan; SC via family runners) tools: mcp__hacker-bob__bob_http_scan, mcp__hacker-bob__bob_http_confirm, mcp__hacker-bob__bob_read_http_audit, mcp__hacker-bob__bob_read_surface_routes, mcp__hacker-bob__bob_read_candidate_claims, mcp__hacker-bob__bob_read_verification_round, mcp__hacker-bob__bob_read_verification_context, mcp__hacker-bob__bob_write_evidence_packs, mcp__ha
Alternatives in Development
- Gitnexus Synthesis Critic — GitNexus final review synthesis critic 45.8k ★
- Finding Checker — Blind adversarial checker — given ONLY a finding artifact and its evidence (never the author's reasoning), tri 348 ★
- Acx Shipper — AgentCortex /ship phase executor 145 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Final Verifier
Round 3 verification — re-runs only REPORTABLE findings with fresh requests as final confirmation
Evaluator Spawn
Generic TaskGraph evaluator shell — executes Transition and Hypothesis nodes the orchestrator dispatched via b
Program Scout
Scouts publicly-listed HackerOne, Bugcrowd, or Intigriti bug-bounty programs likely to yield reportable findin
Theme Ranker
Compares the theme candidates' evidence packs, selects the three core themes, and structures the reasons for s
Incident Debugger
Structured incident debugger. Collects symptoms, generates hypotheses, gathers evidence, and identifies root c
Agentlens Tldr Worker
Token-calibrated worker for ONE bounded code change inside a scan-and-fix run. Navigates with the tldr CLI and
Related Skills
Yotta Compliance
Deterministic, local compliance-clause review skill for AI agents: versioned JSON rule packs, evidence-anchore
Bymax Agent Kit
Engineering toolkit for Claude Code AND Codex, from one source. Bounded dual review (a Claude pass and a Codex
Codex Claude Orchestrator
Codex plugin that delegates a bounded coding task to Claude Code and returns the result to Codex for verificat