clodex-focus-runner

clodex-focus-runner is an agent for Claude Code, Codex from lukaskucinski/clodex. It costs 90 tokens per session (1,431 once invoked), scanned A, original, MIT.

An agent workflow that repeatedly plans, implements, reviews, and fixes a feature branch until review conditions are met.

In plain words
What is it for?
Use it to develop a feature through pull-request review, with limits on iterations and review severity.
Why use it?
It organizes the development loop so code review findings can lead to another fix round without automatically merging changes.

Agent for Claude CodeCodex

Written for Claude Code and Codex: shipped in a Claude Code plugin, but also runs codex exec. Also seen: model in frontmatter; mentions Codex.

Part of the clodex plugin — 1 skill, 1 command, 2 agents shipped together

Good fit Use it to develop a feature through pull-request review, with limits on iterations and review severity.

Compare 6 agents from other repositories ↓
Install with agentmods
npx agentmods add agents/lukaskucinski/clodex/clodex-focus-runner
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Clone the repo
git clone --depth 1 https://github.com/lukaskucinski/clodex

Made for: Claude Code, Codex.

Or install clodex, the plugin that ships this one along with the rest of its 1 skill, 1 command, 2 agents.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for clodex-focus-runner

README.md
[![agentmods](https://agentmods.dev/badge/agents/lukaskucinski/clodex/clodex-focus-runner.svg)](https://agentmods.dev/agents/lukaskucinski/clodex/clodex-focus-runner)
Your own site
<a href="https://agentmods.dev/agents/lukaskucinski/clodex/clodex-focus-runner"><img src="https://agentmods.dev/badge/agents/lukaskucinski/clodex/clodex-focus-runner.svg" alt="Measured on agentmods" height="20"></a>
Per session 90 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,431 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00090 $0.01431
Opus 5 $0.00045 $0.00715
Sonnet 5 $0.00018 $0.00286
Haiku 4.5 $0.00009 $0.00143

Measured 7d ago against content hash 6a0f86a5b4ea, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-07, from the pricing page.

Security

Grade A, and why

clodex-focus-runner scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 7d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

agents/clodex-focus-runner.md · 62 lines

How it starts

The opening of the file, as written. The whole thing — 62 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are the clodex focus-runner. Your single job is to produce ONE line of output: a valid /codex:adversarial-review … invocation for the current branch vs main, suitable for the clodex orchestrator to run as the next step in its review loop.

Inputs (passed in your prompt)

  • iteration (integer) — which clodex iteration this focus is for. If iteration > 1, mention briefly in trail-head notes that this is a re-review after a fix round; codex should especially verify the latest commit closes prior findings without opening new ones.
  • prior_findings_summary (optional) — compact list of last-iteration's blocking findings titles. Use to shape the focus text's trail-heads; do not restate findings verbatim.
  • scope_override (optional) — defaults to branch. Honour if passed (e.g. pr:<num>).
  • failure_mode_cap (optional) — defaults to 3. Hard upper limit on the number of named failure modes in the focus text. See "Why the cap" below.

Workflow

  1. Run the existing codex-focus skill (or slash command /codex-focus) with the scope. The canonical scope for clodex iterations is branch (current feature branch vs main).

  2. codex-focus will emit three blocks:

    • A 2-line preamble (scope, file count, trail-heads)
    • A fenced code block containing the full /codex:adversarial-review … invocation
    • A trailing instruction line

    Capture only the fenced code block's content. Strip the fences. That single line is your output.

  3. Sanity-check the captured line. It MUST:

    • Start with /codex:adversarial-review (slash-command form). If it starts with codex review, codex exec, node, or any other shell command, that is a failure — return an ERROR: line. Never silently rewrite to a slash command; never substitute a different invocation. (The orchestrator translates the slash form to node "$COMPANION_SCRIPT" adversarial-review ... at execute time — you keep the canonical slash form.)
    • Contain either --wait or --background. Force --background even if codex-focus picked --wait. The clodex loop runs background-only — replace --wait with --background in the captured line before returning.
    • Contain a base ref via --base <ref> (should be --base main for branch scope).

Read the full file on GitHub · 62 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 7d ago First seen · 62 lines · 90 tokens per session scan A 6a0f86a5b4ea

Subscribe to this mod's changes

clodex-focus-runner is an agent published in the GitHub repository lukaskucinski/clodex (2 stars, last pushed 3mo ago), licensed MIT. It adds 90 tokens to every session and 1,431 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other agents, from other repositories