A workflow for interactively reviewing findings from an earlier code-review report that an automatic fix would skip. It presents each sufficiently high-scoring finding, its choices, and a recommendation before recording the decision.

Use it to walk through review findings, promote selected items for later handling, skip others, republish the updated report, and add an audit comment to the pull request.

It helps a developer make deliberate decisions on important findings that should not be automatically changed or ignored.

walkthrough — Command for Claude Code by adamjgmiller

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/adamjgmiller/adamsreview

Made for: Claude Code.

Or install adamsreview, the plugin that ships this one along with the rest of its 6 commands, 1 hook.

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.

README.md

[![agentmods](https://agentmods.dev/badge/commands/adamjgmiller/adamsreview/walkthrough/github.svg)](https://agentmods.dev/commands/adamjgmiller/adamsreview/walkthrough)

Your own site

<a href="https://agentmods.dev/commands/adamjgmiller/adamsreview/walkthrough"><img src="https://agentmods.dev/badge/commands/adamjgmiller/adamsreview/walkthrough/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/commands/adamjgmiller/adamsreview/walkthrough"><img src="https://agentmods.dev/badge/commands/adamjgmiller/adamsreview/walkthrough.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 43 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.

When invoked 17,950 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.

Model	Per session	Once invoked
Fable 5.1	$0.00043	$0.17950
Opus 5	$0.00022	$0.08975
Sonnet 5	$0.00009	$0.03590
Haiku 4.5	$0.00004	$0.01795

Measured 10d ago against content hash 82e7a377ebf6, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-09, from the pricing page.

Security

Grade A, and why

walkthrough 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 10d 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.

commands/walkthrough.md · 1,718 lines

How it starts

The opening of the file, as written. The whole thing — 1,718 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Walk the reviewer through findings in the latest /adamsreview:review artifact that /adamsreview:fix would skip — deep-manual, deep-report, light-manual, light-report, and light-auto that fails the impact_type lane filter — further restricted to findings scoring at or above $threshold so low-signal items don't pad the session. For each finding, dispatch a Sonnet briefing agent (claim → options → recommendation), ask the reviewer to decide, and record a promote (via the shared promote-core.md fragment with --defer-publish semantics) or a skip. At the end, render + publish the updated review once, then POST a separate decisions-log comment to the PR for audit.

Read fragments/_prelude-shared.md before proceeding — it lists rules that apply to every step below (sub-agent return handling, helper-script error-as-prompt).

Arguments

[threshold] (optional, positional) — non-negative integer score floor. Default: 60. Findings with effective score (COALESCE(score_phase4, score_phase3, -1)) below this value are dropped from the walk scope so the session isn't padded with low-signal findings. Independent of the /adamsreview:fix threshold — promoted findings are picked up by /adamsreview:fix regardless of the score gate via the human_confirmation bypass.

What it does

Parses the threshold.
Locates the artifact for the current branch.
Computes three walkthrough scopes (qualifying / full / pre-existing).
Shows a pre-flight summary + asks the reviewer which tier to walk (default Qualifying; Full skip set is the opt-in for auditing Phase-3-demoted findings). 4.5. Auto-recommendation batch. When any in-scope finding carries an auto_fix_hint from Phase 5.5, surfaces a batch confirm UI (Accept all / Pick subset / Walk through each). Accepted findings are promoted via --apply-auto-rec-promotions and removed from the per-finding loop's working scope.
For each remaining finding in the chosen tier:
- When the finding has auto_fix_hint (i.e., it was in scope at §4.5 but the reviewer chose Walk-each / didn't include it in Pick-subset), constructs the briefing inline from auto_fix_hint — saves the ~3-5k tokens a Sonnet briefer would cost.
- Otherwise, dispatches a Sonnet briefing sub-agent → {summary, options[], recommendation}.
- Presents the briefing and asks the reviewer which option to pick (or "Edit the fix hint" to override the recommended option's hint).
- Dispatches a promote (patch + trace, no render/publish) or a skip.
Re-renders artifact.md once.
Re-publishes the main review comment once.
For each pre_existing_report finding (PR mode only): offers to draft + create a GitHub issue, one by one.
Posts a new "Walkthrough decisions" comment to the PR with the full log of what was auto-accepted / promoted / skipped / issues filed / why.
Appends a ## walkthrough (<ts>) block to trace.md.
Prints a user-visible summary.

Read the full file on GitHub · 1,718 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.

10d ago First seen · 1,718 lines · 43 tokens per session scan A 82e7a377ebf6

Subscribe to this mod's changes

walkthrough is a command published in the GitHub repository adamjgmiller/adamsreview (243 stars, last pushed 3mo ago), licensed MIT. It adds 43 tokens to every session and 17,950 once invoked, about $0.0002 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

walkthrough

Getting it into your agent

Wrote this? Show the measurements

What it costs to keep this loaded

Grade A, and why

Nothing flagged

How it starts

Arguments

What it does

What this file has done since we first saw it

Other commands, from other repositories

sdd-init

review-branch

init

merge-conflict-analysis

repo-audit

argos