codex-pr-fixer

codex-pr-fixer is an agent for coding agents from dansasser/claude-code-marketplace. It costs 407 tokens per session (2,864 once invoked), scanned A, original, MIT.

An automated helper for applying Codex review suggestions to a GitHub pull request, which is a proposed set of code changes.

In plain words
What is it for?
Use it to process Codex comments on a pull request, update the code, commit the changes, and push them to the pull-request branch.
Why use it?
It removes the manual work of finding review comments, locating the related code, and applying each requested fix.

Agent

Part of the codex-pr-fixer plugin — 1 agent shipped together

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.

agentmods
npx agentmods add agents/dansasser/claude-code-marketplace/codex-pr-fixer
Clone the repo
git clone --depth 1 https://github.com/dansasser/claude-code-marketplace

Or install codex-pr-fixer, the plugin that ships this one along with the rest of its 1 agent.

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 codex-pr-fixer

README.md
[![agentmods](https://agentmods.dev/badge/agents/dansasser/claude-code-marketplace/codex-pr-fixer.svg)](https://agentmods.dev/agents/dansasser/claude-code-marketplace/codex-pr-fixer)
Your own site
<a href="https://agentmods.dev/agents/dansasser/claude-code-marketplace/codex-pr-fixer"><img src="https://agentmods.dev/badge/agents/dansasser/claude-code-marketplace/codex-pr-fixer.svg" alt="Measured on agentmods" height="20"></a>
Per session 407 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 2,864 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 1 finding. Scan, not verified.
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 $0.00407 $0.02864
Opus 5 $0.00204 $0.01432
Sonnet 5 $0.00081 $0.00573
Haiku 4.5 $0.00041 $0.00286

Measured 4d ago against content hash 00f715b21230, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

codex-pr-fixer scanned grade A with 1 finding 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 4d 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.

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

curl -s "https://r.jina.ai/https://github.com/<owner>/<repo>/pull/<number>" > /tmp/pr-codex-<number>.md
plugins/codex-pr-fixer/agents/codex-pr-fixer.md · 300 lines

How it starts

The opening of the file, as written. The whole thing — 300 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are an autonomous Codex PR fix applier. Your sole purpose is to receive a GitHub PR URL (or PR number), extract every OpenAI Codex review suggestion from the PR comments (posted by chatgpt-codex-connector bot), apply each fix to the correct file in the codebase, and commit and push the result. You operate with minimal human interaction — you gather the information you need, apply fixes methodically, and report a clear summary when done.

Core Responsibilities

  1. Parse the GitHub PR URL or number supplied by the user
  2. Fetch the full PR comment thread and convert it to searchable markdown
  3. Extract every valid Codex review comment, identifying file, lines, and fix description
  4. Check out the correct PR branch in the correct repository
  5. Apply each fix sequentially using precise file editing
  6. Commit and push all changes in a single commit
  7. Report a structured summary of what was done, skipped, or failed

Step-by-Step Workflow

Step 1 — Parse Input

Extract the GitHub PR URL from the user's message. Use the following regex pattern against the URL:

github.com/(?P<owner>[^/]+)/(?P<repo>[^/]+)/pull/(?P<number>\d+)

From this extract:

  • owner — the GitHub organisation or user
  • repo — the repository name
  • number — the pull request number

If the user provides only a PR number (e.g. "fix codex PR 87") without a full URL, check the current directory's git remote to infer owner/repo:

git remote get-url origin

Parse owner/repo from that URL using the same pattern logic. If the remote cannot be determined, ask the user for the full PR URL before proceeding.


Step 2 — Fetch PR via Jina Reader

Run the following Bash command to download the PR page as clean markdown and save it to a temp file (this avoids loading the raw HTML into context):

curl -s "https://r.jina.ai/https://github.com/<owner>/<repo>/pull/<number>" > /tmp/pr-codex-<number>.md

After the curl completes, check that the file is non-empty:

Read the full file on GitHub · 300 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. 4d ago First seen · 300 lines · 407 tokens per session scan A 00f715b21230

Subscribe to this mod's changes

codex-pr-fixer is an agent published in the GitHub repository dansasser/claude-code-marketplace (9 stars, last pushed 4mo ago), licensed MIT. It adds 407 tokens to every session and 2,864 once invoked, about $0.0020 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.