Verification Runner
Run project-aware verification loop. Reads mix.exs to discover tools (credo, dialyzer, sobelow, ex_check), test commands, and custom aliases. Use proactively after code changes.
- Type
- Subagent
- Repository
- oliver-kriska/claude-elixir-phoenix
- GitHub stars
- 556
- License
- MIT
- Repo last updated
- Sep 25, 2026
- Model
- haiku
What Verification Runner is
Verification Runner is a subagent published in the oliver-kriska/claude-elixir-phoenix repository on GitHub, which has about 556 stars. The repository describes itself as: “Claude Code plugin for Elixir/Phoenix/LiveView — 26 specialist agents, Iron Laws enforcement, and Tidewave MCP integration. Plan features with parallel research agents, execute with automatic verification, review with 4-agent parallel audits, and capture learnings as reusable knowledge.”
A subagent is a specialist assistant that Claude can hand part of a task to. It is a markdown file whose frontmatter sets a name, a description that tells Claude when to delegate, and optionally the tools and model it may use; the body becomes the subagent's own system prompt.
Because a subagent works in its own context, it keeps the main conversation focused: Claude can send a narrow job, such as a review or a specialised analysis, to Verification Runner and get back a compact result.
It is set up to use these tools: Read, Grep, Glob, Bash, Write. Limiting tools is a good sign: the subagent can only do what those tools allow.
How to install Verification Runner
Claude Code
- Download verification-runner.md from the repository.
- Save it to ~/.claude/agents/ to use it in every project, or to .claude/agents/ inside one project to share it through version control.
- Claude Code watches these folders, so the subagent is usually available right away. Ask Claude to use it by name, or @-mention it to make sure it runs.
Claude Cowork
- Cowork loads subagents through plugins. If the repository is packaged as a plugin marketplace, add it under Customize → Plugins → Add marketplace and install the plugin that contains this subagent.
- Otherwise, bundle the file into your own plugin's agents/ folder and upload it from Customize → Plugins.
New to extending Cowork? Our plugins guide and Customize guide explain how skills, plugins, and connectors fit together.
Inside the source file
An excerpt from plugins/elixir-phoenix/agents/verification-runner.md, shared under the repository's MIT license. Read the full file on GitHub.
You run a project-aware Elixir/Phoenix verification loop. Always discover what the project has before running checks. After core verification passes, offer additional test commands the project has available.
CRITICAL: Compile First, Save Findings File Early
Your orchestrator reads results from the exact file path given in the prompt (e.g., .claude/plans/{slug}/reviews/verification.md). The file IS the real output — your chat response body should be ≤300 words.
Turn budget rules (you have only 15 turns):
- Turn 1: start compiling immediately. Your FIRST Bash call combines discovery and compile so compilation is never deferred: cat mix.exs && mix compile --warnings-as-errors 2>&1 | tail -40 Large projects compile slowly — kicking it off first means you never exhaust turns waiting on it. Use a generous Bash timeout; do NOT poll compilation with repeated "check again" calls.
- Next ~8 turns: remaining verification commands
- By turn ~12: call Write with the verification report — do NOT wait. A partial report beats no file when turns run out.
- If the prompt does NOT include an output path, default to .claude/reviews/verification.md.
You have Write for your own report ONLY. Edit and NotebookEdit are disallowed — you cannot modify source code.
Step 0: Project Discovery (MANDATORY)
Read mix.exs and extract:
- Dependencies — search deps for:
- :credo, :dialyxir, :sobelow, :ex_check, :excoveralls, :boundary
- E2E deps: :phoenix_test_playwright, :phoenix_test, :wallaby
- Aliases — categorize ALL test-related aliases:
- Composite verify: ci:, check:, precommit:
- Unit test variants: test:, "test.with_coverage":, "test.ci":
- E2E test: "playwright.test":, "playwright.run":, "cypress.run":
- Map each alias to which steps it covers
- CLI config — check cli/0 for preferred_envs (newer Elixir) or project/0 [:preferred_cli_env] (older). Note custom MIX_ENV per command.
- ex_check config — if :ex_check in deps, read .check.exs for the full tool pipeline. mix check replaces individual steps.
Report discovery:
Project tools: compile ✓ | format ✓ | credo ✓/✗ | dialyzer ✓/✗ | sobelow ✓/✗ | ex_check ✓/✗
Test commands: mix test (unit) | mix playwright.test (E2E, MIX_ENV=int_test)
Composite runner: mix check (.check.exs) — or "none found"
Strategy: {what will be run}Verification Sequence
Priority 1: ex_check
If .check.exs exists: mix check 2>&1. Skip to Step 7.
Priority 2: Composite alias
If mix ci or similar: run it, then uncovered steps. Skip to Step 7.
Priority 3: Individual steps
- mix compile --warnings-as-errors 2>&1 — always (Elixir 1.20+/OTP 27+: the compiler's built-in type checker surfaces type violations / verified bugs here — --warnings-as-errors fails on them, no Dialyzer required)
- mix format --check-formatted 2>&1 — always
- mix credo --strict 2>&1 — if installed
- mix test --trace 2>&1 — use project alias if exists
- mix dialyzer 2>&1 — if installed, pre-PR
- mix sobelow --config 2>&1 — if installed
Skip unavailable tools: "Credo: ⏭ Not installed"
Step 7: Additional Test Offer
After core passes, list discovered additional test commands:
Core verification passed. Additional test commands available:
1. mix playwright.test (E2E, MIX_ENV=int_test) — full setup + tests
2. mix playwright.run (E2E fast — skips setup)
3. mix test.with_coverage (unit + coverage report)
Run any of these? [1/2/3/all/skip]Use correct MIX_ENV from preferred_envs for each command.
Output Format
# Verification Report
## Project Config
{discovery summary}
## Summary
| Step | Status | Details |
|------|--------|---------|
| Compile | ✅/❌ | {details} |
| Format | ✅/❌ | {details} |
| Credo | ✅/❌/⏭ | {details or "not installed"} |
| Test | ✅/❌ | {pass/fail count} |
| Dialyzer | ✅/❌/⏭ | {details or "not installed"} |
| Sobelow | ✅/❌/⏭ | {details or "not installed"} |
## Overall: ✅ PASS / ❌ FAIL
… Before you install
- Read the whole file first. Skills, commands, and subagents are instructions Claude will follow, so make sure they match what you want.
- Check which tools, scripts, or MCP servers it uses. Local servers and scripts run with your permissions.
- Try it in a test project or a copy of your files before pointing it at real work.
- Pin the version you tested, and review changes before updating.
- Watch for instructions that fetch web content or run shell commands; those are where prompt injection risks start. See our prompt injection guide.
FAQ
What is Verification Runner?
Verification Runner is a subagent for Claude Code and Claude Cowork from the oliver-kriska/claude-elixir-phoenix repository on GitHub. Run project-aware verification loop. Reads mix.exs to discover tools (credo, dialyzer, sobelow, ex_check), test commands, and custom aliases. Use proactively after code changes.
How do I install Verification Runner in Claude Code?
Download verification-runner.md from the repository. Save it to ~/.claude/agents/ to use it in every project, or to .claude/agents/ inside one project to share it through version control. Claude Code watches these folders, so the subagent is usually available right away. Ask Claude to use it by name, or @-mention it to make sure it runs.
Can I use Verification Runner in Claude Cowork?
Cowork loads subagents through plugins. If the repository is packaged as a plugin marketplace, add it under Customize → Plugins → Add marketplace and install the plugin that contains this subagent. Otherwise, bundle the file into your own plugin's agents/ folder and upload it from Customize → Plugins.
Is Verification Runner safe to install?
It is a third-party community resource, not reviewed by Anthropic or this site. Read the source file first, check which tools and connectors it uses, and install only from sources you trust.
Similar resources
- Ash Policy Reviewer Ash policy security reviewer — audits policies, checks, and authorization rules for gaps, bypass patterns, and ordering hazards. Use proactively on Ash resources with policies do blocks or checks/ modules. Subagent · oliver-kriska/claude-elixir-phoenix
- Call Tracer Orchestrates parallel call tree tracing using subagents for each entry point category (Controllers, LiveViews, Workers, GenServers). Use proactively when debugging unexpected values, tracing request flow, or planning signature changes. Subagent · oliver-kriska/claude-elixir-phoenix
- Ash Query Optimizer Ash query optimizer — detects N+1 loads, suggests aggregates over load+Enum, identifies calculation vs load tradeoffs. Use when reviewing Ash queries, LiveView data loading, or domain action efficiency. Subagent · oliver-kriska/claude-elixir-phoenix
- Catchup Async-team return-from-absence briefing. /catchup fans out to GitHub, git, Linear, and calendar, then emits one prioritized Context Brief instead of a firehose. Plugin · oliver-kriska/claude-elixir-phoenix
- Web Researcher Fetches and extracts information from web sources efficiently. Optimized for ElixirForum, HexDocs, and GitHub. Spawned by /phx:research or planning-orchestrator with pre-searched URLs or focused queries. Subagent · oliver-kriska/claude-elixir-phoenix
- Testing Reviewer Reviews test code for Elixir best practices - ExUnit patterns, Mox usage, LiveView testing, factory patterns. Use proactively after writing tests or during code review. Subagent · oliver-kriska/claude-elixir-phoenix
- Workflow Orchestrator Orchestrates the full agentic workflow cycle (plan → work → review). Internal use by /phx:full command. Subagent · oliver-kriska/claude-elixir-phoenix
- Skill Effectiveness Analyzer Analyzes skill effectiveness data to identify failure patterns and recommend improvements. Use after /skill-monitor flags underperforming skills. Subagent · oliver-kriska/claude-elixir-phoenix