# Test Agent

> Skill: TestDefinitionAgent (TDD-First, Evidence-Gated)

- Skill: `farmountain/test-agent` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add farmountain/test-agent`
- Raw SKILL.md: https://api.skillmd.com/api/skills/farmountain/test-agent/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: farmountain (https://skillmd.com/u/farmountain)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/farmountain/test-agent

---

# Skill: TestDefinitionAgent (TDD-First, Evidence-Gated)

## Purpose
Define tests BEFORE execution. Tests here are evidence requirements.
Outputs become EGD entries and verifier checklists.

## Inputs
- SPEC Card
- Mode
- Constraints
- Quality gates (capabilities/quality_gates.md)

## Output (MANDATORY): TEST Card
### TEST Card
- Test objectives:
- Evidence required (list of repo artifacts, logs, checklists):
- Verification method (how verifier will check):
- Negative tests / falsifiers (what would prove failure):
- Risk-triggered approval tests (when Decision Card is mandatory):

## Rules
- Must be verifiable using repo artifacts (no vague claims).
- Must explicitly state "what evidence must exist for done".
- No execution plan allowed until TEST Card exists.