A command that runs an implementation review for the current coding workflow and records responses to its findings. It can also approve the implementation phase when the workflow is waiting for approval.

Running an implementation review, responding to each finding as fixed, rejected, or intentionally left unresolved, and then continuing the review process.

It organizes review feedback and prevents findings from being left unanswered before the workflow continues.

impl — Command by boyand

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 commands/boyand/codex-review/impl

Clone the repo

git clone --depth 1 https://github.com/boyand/codex-review

Per session 12 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.

When invoked 568 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. 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.

Model	Per session	Once invoked
Fable 5	$0.00012	$0.00568
Opus 5	$0.00006	$0.00284
Sonnet 5	$0.00002	$0.00114
Haiku 4.5	$0.00001	$0.00057

Measured yesterday against content hash e0795c703417, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

impl 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 yesterday.

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.

commands/impl.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.

/codex-review:impl

Run an implementation review round immediately for the current Claude thread's active workflow.

If the current workflow is still in plan but already at the approval step, this command should treat the impl invocation itself as explicit approval, advance the workflow to implement, and continue without requiring a separate approval command.

The user may pass:

focus text
--workflow <id> to target a specific workflow

Required behavior

Run the engine:

"${CLAUDE_PLUGIN_ROOT}/.bin/codex-review" impl $ARGUMENTS

Do not pass hidden runtime flags like --session-pid or --session-id; the wrapper injects them automatically.
If the plan phase is already at approval, do not stop to ask for a separate approval command first. Running impl is the approval signal.

If the engine reports a review file, read it completely.
Write findings responses automatically to the reported Findings file:

respond to every finding ID
use FIX, REJECT, or WONTFIX
include concrete justification for any REJECT or WONTFIX

After writing findings:

run:

"${CLAUDE_PLUGIN_ROOT}/.bin/codex-review" summary

present that summary output as the approval summary for the round
in this same command response only, ask the user inline what they want to do next:
- finish the workflow
- run another implementation review round
- stop and keep current results
do not re-offer these choices on unrelated later turns

if the user chooses to finish the workflow, run:

"${CLAUDE_PLUGIN_ROOT}/.bin/codex-review" __approve

if the user chooses another implementation review round, run:
```
"${CLAUDE_PLUGIN_ROOT}/.bin/codex-review" __repeat
```
if the user chooses to stop and keep current results, run:
```
"${CLAUDE_PLUGIN_ROOT}/.bin/codex-review" __done
```
do not require the user to type separate slash commands for those transitions

Rules:

Do not manually edit workflow.json or decisions.tsv.
Do not hand-roll the post-review counts; use the summary command output after writing findings.
Do not ask the user to run a separate respond command.

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.

yesterday First seen · 62 lines · 12 tokens per session scan A e0795c703417

Subscribe to this mod's changes

impl is a command published in the GitHub repository boyand/codex-review (16 stars, last pushed 16d ago), licensed MIT. It adds 12 tokens to every session and 568 once invoked, about $0.0001 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 2026-08-30.

Code review Productivity

impl

Getting it into your agent

What it costs to keep this loaded

Grade A, and why

Nothing flagged

How it starts

/codex-review:impl

Required behavior

What this file has done since we first saw it

Other commands, from other repositories

checklist

clarify

specify

analyze

converge

implement