smithery/gittower

validate-tests

Codex-gate the test plan in an implementation plan before implementation starts

Installation

$ npx skills add smithery/gittower --skill validate-tests

Similar popular skills

Related neighbors and high-traction skills in the same topics — useful to compare before installing.

Also in this package

Other skills from smithery/gittower.

npx skills add smithery/gittower

Browse all from smithery/gittower

More details

Agent compatibility

Declared targets from SKILL.md / docs. Unmarked agents are not listed — the skill may still install via the CLI.

Claude Code Not declared
Cursor Not declared
Codex Not declared
GitHub Copilot Not declared
Windsurf Not declared
Gemini CLI Not declared
Cline Not declared
OpenCode Not declared

Skill metadata

Parsed from SKILL.md frontmatter.

Allowed toolsRead, Grep, Glob, Edit, Write, Bash, Agent

Package contents

Files included with this skill beyond the listing page.

  • skill md SKILL.md 2,648 B
  • docs SUMMARY.md 86 B

History

  1. First recorded snapshot · 0 installs

SKILL.md

Validate Tests (Codex Gate)

Gate the test plan of an implementation plan before any implementation starts. Combines a local check against the testing guidelines with an external Codex review, following the shared convention in [../shared/CODEXGATE.md](../shared/CODEXGATE.md).

This is a required step between /create-plan and /implement. Once the plan passes this gate, the test plan is authoritative — /implement treats it as immutable.

Instructions

  1. Find the Plan

- Detect current workflow folder from git branch - Read .ai/<folder>/plan.md - If no plan exists, suggest running /create-plan first

  1. Local Validation

Load TESTINGGUIDELINES.md and GITTEST_SCENARIOS.md, then check the Test Plan section: - Each scenario follows the guidelines (setup patterns, assertions that verify behavior rather than just "no error", proper Git scenario setup) - Pay special attention to rules marked CRITICAL in the guidelines - All code paths, error conditions, and edge cases have scenarios - Scenarios are concrete enough to implement without guessing

Fix what you find directly in the plan.

  1. Codex Gate

Run a Codex gate per [../shared/CODEXGATE.md](../shared/CODEXGATE.md):

- Artifact: the Test Plan section of .ai/<folder>/plan.md - Task for Codex: find missing scenarios, wrong or underspecified expected outcomes, and scenarios that don't match the source of truth. TDD framing: these tests define the design — gaps found now are cheap, gaps found during implementation are expensive - Source of truth: the spec issue or analysis.md/concept.md the plan was created from (pass its content or path) - Guidelines: TESTINGGUIDELINES.md, GITTEST_SCENARIOS.md

  1. Evaluate and Apply

Evaluate every Codex finding per the convention: apply only with high confidence, reject over-engineering and hypothetical scenarios, when in doubt leave it out. Update the plan's Test Plan section with applied findings.

Log all findings — applied, partial, and rejected with reasons — to .ai/<folder>/codex-test-plan.md.

  1. Report Findings

Summarize: - Local validation: issues found and fixed - Codex gate: findings applied / rejected (with the log path) - Open questions that need the user (genuine design decisions only) - Confirmation that the plan is gated and ready for /implement