Spec Runner — Testing & QA agent for Claude Code
Runs the Savi specs affected by the current change and triages every failure into real-bug / stale-spec / pre-existing / environment / vacuous.
How to install Spec Runner
Installs to ~/.claude/agents/below9k-claude-skills-spec-runner.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/below9k/claude-skills/HEAD/agents/spec-runner.md -o ~/.claude/agents/below9k-claude-skills-spec-runner.md Restart Claude Code, or start a new session, for it to be picked up.
What Spec Runner does
name: spec-runner description: Runs the Savi specs affected by the current change and triages every failure into real-bug / stale-spec / pre-existing / environment / vacuous. Detects the Savi line (DSP or main line) and uses the runner paths that actually work there; never improvises a runner. Use after edits, before a commit, or when a spec fails and you need to know whether to believe it. tools: Bash, Read, Grep, Glob model: sonnet
Spec runner
Detect the Savi line first
Run at
Alternatives in Testing & QA
- Bash32 Portability Reviewer — Reviews Mole shell and Bats diffs against the current macOS Bash 3.2, errexit, timeout, TTY, BSD-tool, and CI 72k ★
- 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 ★
- Backend Unit Tests — Add focused backend unit tests aligned with current package test conventions 720 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Fe Test Runner
Dedicated test execution. Returns only a categorized summary, keeping failure stack traces out of the parent c
CI Failure Responder
Triages and fixes CI failures for the devloop project. Given a branch name or gh run URL, pulls failure output
Baseline Verifier
Subagent dispatched by do-test to classify test failures as regressions vs pre-existing by running failing tes
Develop
Implements a fix or feature for a single GitHub issue in an isolated worktree. Iterates up to 3 cycles (merge
Test Verifier
Runs and verifies test cases from tests/test.md. Prefers e2e testing with curl/bash. Looks for environment inf
AI Tester
Runs the project's tests for a change and classifies the outcome as PASS, EXISTING TEST FAILURE, NEW REGRESSIO
Related Skills
Cve Env
Agentic CVE → Docker environment builder: given a CVE ID, builds and verifies a Docker environment running the
E2e Auto
Unattended E2E sweep. Detects stack (flutter web), boots app, plans flows, dispatches per-flow runners in para
Claude Workflows
Claude Code plugin: two workflow skills layered on Superpowers. fixing-raw-bugs triages unstructured bug repor