Plan Hv Agent — Development agent for Claude Code
Browser-verification owner for closing observable plan gates with fresh evidence.
How to install Plan Hv Agent
Installs to ~/.claude/agents/startmeupai-swe-agents-plan-hv-agent.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/startmeupai/swe-agents/HEAD/.claude/agents/plan-hv-agent.md -o ~/.claude/agents/startmeupai-swe-agents-plan-hv-agent.md Restart Claude Code, or start a new session, for it to be picked up.
What Plan Hv Agent does
name: plan-hv-agent description: Browser-verification owner for closing observable plan gates with fresh evidence. skills:
- plan-hv-automation
- playwright-testing
Plan Human-Verification Agent
Purpose and Responsibility
Triage open plan gates, close browser-observable ones, and return only genuine human/provider decisions.
When to Use
Use when an implemented plan has open runtime or human-verification gates.
Inputs
- Plan, gate criteria, profiles, environment, fix
Alternatives in Development
- Dashclaw Gate Runner — Runs the DashClaw verification gates (lint, full vitest suite, build, contract checks) and returns ONLY the fa 297 ★
- Acx Implementer — AgentCortex /implement phase executor 145 ★
- Ac Evidence Verifier — Phase 3.4 evidence verifier 115 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Evidence Collector
Independent verification agent. Takes FINAL-PLAN.md and independently verifies every acceptance criterion was
Playwright Investigator Agent
Browser-debugging owner for reproducing defects and fixing application causes.
Harness Shipper
Fresh shipping agent for goal loops. Runs the no-mistakes validation pipeline exactly once after Checker PASS,
Run Issues Implementer Escalated
Third-attempt implementer for a /run-issues issue that two previous implementers failed to get past the gates.
Idea Scorer
Gates candidate ideas on Stage One, then scores survivors against the real rubric. Runs in a fresh context dur
Loop Judge
Fresh-context cold-read judge for subjective acceptance criteria (readability, API clarity). Runs after mechan
Related Skills
J Execute Plan
Execute a written implementation plan task-by-task with verification gates — inline batches or a fresh subagen
Autonomy Off
End autonomy mode — finish the in-flight slice to merge, start nothing new, then post a closing summary (merge
Gaia Debt
Fix the tech-debt backlog, a single issue or a recommended related batch, highest severity then oldest first,