Baseline Verifier — Testing & QA agent for Claude Code
Subagent dispatched by do-test to classify test failures as regressions vs pre-existing by running failing tests against main.
How to install Baseline Verifier
Installs to ~/.claude/agents/tomcounsell-ai-baseline-verifier.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/tomcounsell/ai/HEAD/.claude/agents/baseline-verifier.md -o ~/.claude/agents/tomcounsell-ai-baseline-verifier.md Restart Claude Code, or start a new session, for it to be picked up.
What Baseline Verifier does
name: baseline-verifier description: "Subagent dispatched by do-test to classify test failures as regressions vs pre-existing by running failing tests against main. Returns structured JSON classification." color: yellow tools: ['*']
Baseline Verifier
Purpose
You are a **test failure classifier**. You receive a list of failing test node IDs from a feature branch, run those same tests against `main`, and classify each failure as a **regression**, **pre-existing**, or **inconclusive
Alternatives in Testing & QA
- Civitai Test Review — Reviews the tests in a feature segment of the main Civitai Next.js app (src/) for whether they would actually 7.2k ★
- Browser Verifier — Uses Playwright MCP to smoke-test the running application in a browser 63 ★
- Gsd Doc Classifier — Classifies a single planning document as ADR, PRD, SPEC, DOC, or UNKNOWN 57 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Spec Runner
Runs the Savi specs affected by the current change and triages every failure into real-bug / stale-spec / pre-
Waves Implementer
Implements one task of one waves plan inside an existing worktree — failing test first, gate green, one commit
E2e Failure Classifier
Use when a Playwright or Cypress test failure needs a root cause, to classify ONE failure into the F1–F15 taxo
Forge Visual Verifier
Perceptual gate for spec [visual] acceptance criteria. Drives Playwright MCP (navigate + take_screenshot + eva
05 Defect Triage.Agent
Defect classification and triage specialist. Takes raw failures from test execution and produces: severity/pri
Skill Verifier
Fresh-session surrogate verifier for improve skill rewrites. Use when /improve reaches Step 4.5 and needs isol
Related Skills
Pharn Dev Regress
Detect regressions OUTSIDE the just-built feature: re-run the existing deterministic suite (npm run check's ga
Classify Backfill
Backfill classification data from existing Claude Code session logs for a project
Respond Reviewer
Classify a reviewer letter into per-comment labels and dispatch each to the right specialist; assemble a rebut