testerarmy-cli

Installation
SKILL.md

TesterArmy CLI

Validation workflow for repositories using ta / testerarmy.

When to Use

  • After any product code change before marking work done
  • When reproducing UI/API regressions with browser-level checks
  • When validating auth, onboarding, dashboard, project, team, billing, or settings flows
  • When you need machine-readable QA output for CI or artifacts

Core Commands

# Check auth/key state first (recommended)
ta status --json

# Authenticate only if needed (or set TESTERARMY_API_KEY)
ta auth

# Run a single markdown scenario (human quick check)
ta run tests/01-landing-page.md

# Run all top-level tests in directory (parallel, default 3)
ta run tests/ --json

# Override concurrency
ta run tests/ --json --parallel 5

# Inspect artifacts
ta run tests/02-quick-test-runner.md --json

# Structured output for automation
ta run tests/03-create-project.md --json --output .testerarmy/latest-run.json

# List recent local runs
ta list

# Same as list (short alias)
ta ls

Standard Validation Flow

Agent Defaults

  • Agents should default to ta run ... --json
  • Add --output <file> when downstream steps need to read the result
  • Plain text output is fine for human-only quick local checks
  1. Check auth/key status first:
ta status --json

If authenticated is false, run ta auth (or set TESTERARMY_API_KEY).

  1. Start app under test:
pnpm dev
  1. Point tests to target app:
export TESTERARMY_TARGET_URL="http://localhost:3000"
  1. Run at least one relevant validation scenario, either:
  • a targeted markdown test from tests/
  • an ad hoc prompt with ta run "..." --json
  1. For cross-cutting changes, run broader coverage:
ta run tests/ --json --parallel 3
  1. Report exact validation command(s) and result in final update.

Scenario Selection

  • Landing/public pages: tests/01-landing-page.md
  • Quick test runner: tests/02-quick-test-runner.md
  • Project flows: tests/03-create-project.md, tests/04-project-overview.md, tests/11-project-settings.md
  • Run history: tests/05-runs-history.md
  • Team/settings/auth areas: tests/07-settings-check.md, tests/08-create-team.md, tests/09-invite-member.md, tests/10-create-api-key.md
  • Billing: tests/12-billing-checks.md

If no existing scenario covers your change, run a focused prompt with explicit URL:

ta run "verify <changed behavior>" --url "$TESTERARMY_TARGET_URL" --json

Ad hoc prompts are valid for final validation, not only tests/*.md files.

Test Prompt Composition Rules

  • Prefer markdown tests over ad-hoc prompts when available
  • Keep shared auth/setup in TESTER.md (CLI auto-prepends it for directory and file runs)
  • Do not hardcode environment-specific URLs inside test files; use TESTERARMY_TARGET_URL

Artifacts and Debugging

Each run writes to .testerarmy/<timestamp>/:

  • run-meta.json: lifecycle metadata (pid, timestamps, prompt, target URL)
  • result.json: normalized QA outcome (PASS / FAILED, issues, description)
  • debug-run.json: full stream + tool timeline when available

Useful triage flow:

ta list
ta run tests/01-landing-page.md --json --headed --timeout 900000

For multiple concurrent/background runs, use ta ls to check which runs are still RUNNING and which finished with PASS / FAILED.

Exit Codes

  • 0: all tests passed
  • 1: one or more tests failed (or run cancelled)
  • 2: runtime/CLI error

Treat non-zero as validation failure.

Required Reporting Pattern

In final updates after code changes:

  • Include the ta run ... command(s) used
  • Include pass/fail outcome
  • Prefer the --json command variant in agent reports
  • If failed, include what failed and artifact path under .testerarmy/

Do not claim work is validated without a TesterArmy CLI run.

References

File Description
reporting-template.md Copy-paste format for final validation status updates
Installs
9
Repository
tester-army/cli
GitHub Stars
29
First Seen
Mar 10, 2026