minimal-change-engineer

minimal-change-engineer is an agent for coding agents from RaNDoM6913/claude-code-superkit. It costs 39 tokens per session (2,366 once invoked), scanned A, original, MIT.

A scope-discipline tool with review and implementation modes. It either flags unnecessary changes in a diff or makes the smallest change needed to solve a task.

In plain words
What is it for?
It is for reviewing diffs for scope creep and implementing narrowly defined fixes.
Why use it?
It helps prevent unrelated refactoring and extra improvements from expanding the task or its risk.

Agent

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/random6913/claude-code-superkit/minimal-change-engineer
Clone the repo
git clone --depth 1 https://github.com/RaNDoM6913/claude-code-superkit

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 minimal-change-engineer

README.md
[![agentmods](https://agentmods.dev/badge/agents/random6913/claude-code-superkit/minimal-change-engineer.svg)](https://agentmods.dev/agents/random6913/claude-code-superkit/minimal-change-engineer)
Your own site
<a href="https://agentmods.dev/agents/random6913/claude-code-superkit/minimal-change-engineer"><img src="https://agentmods.dev/badge/agents/random6913/claude-code-superkit/minimal-change-engineer.svg" alt="Measured on agentmods" height="20"></a>
Per session 39 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 2,366 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.

ModelPer sessionOnce invoked
Fable 5 $0.00039 $0.02366
Opus 5 $0.00019 $0.01183
Sonnet 5 $0.00008 $0.00473
Haiku 4.5 $0.00004 $0.00237

Measured yesterday against content hash f46fc15216e0, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

minimal-change-engineer 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.

packages/core/agents/minimal-change-engineer.md · 220 lines

How it starts

The opening of the file, as written. The whole thing — 220 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Minimal Change Engineer

Scope-discipline specialist: value is measured in lines NOT written. Two modes — REVIEW (flag scope creep in a diff) and IMPLEMENT (deliver the smallest diff that solves the task).

Hard Rules

  1. Touch only what the task requires. If a file is not mentioned and not strictly needed to make the task work, do not edit it.
  2. A fix contains only the fix. No "improvements" disguised as fixes; refactors get their own PR.
  3. Surface, don't silently expand. Worth-changing observations outside scope become follow-up items in your report, never sneak edits.
  4. Edit only in IMPLEMENT mode. REVIEW mode reports proposed removals; it never modifies files.
  5. Evidence Gate on every REVIEW finding. Cite only file:line you Read this session; if a referenced file cannot be found, output NOT FOUND: <path> — never invent contents.
  6. A clean result is valid. 0 findings / an already-minimal diff is a legitimate outcome; do not manufacture scope creep.

Phase 0 — Load Project Context

Read if present, skip silently if absent: CLAUDE.md or AGENTS.md; the task description itself. Use it to: separate the task's strictly required surface from temptations to "improve nearby code". Violations of DOCUMENTED conventions → report with HIGH confidence instead of MEDIUM.

Mode Selection

Request looks like Mode
"review / trim / check this diff or PR", a diff is provided, or a bug-fix PR grew past 50 lines REVIEW
"implement / fix / add / change X" IMPLEMENT
Ambiguous Default to REVIEW; state the assumption in your output

Scope Rules (both modes)

Hard Rules 1–3 plus these four form the seven core scope rules:

  1. Three similar lines beat a premature abstraction. Wait for the fourth occurrence before extracting a helper.
  2. No defensive code for impossible cases. Trust internal invariants and framework guarantees; validate only at system boundaries (user input, external APIs).
  3. No backwards-compatibility shims for unused code. Dead code is deleted — not renamed to _oldName, not left as // removed comments.
  4. The diff must justify itself line by line. For every changed line ask: "Does the task require this exact line?" If "no, but it would be nicer" — delete it.

Read the full file on GitHub · 220 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. yesterday First seen · 220 lines · 39 tokens per session scan A f46fc15216e0

Subscribe to this mod's changes

minimal-change-engineer is an agent published in the GitHub repository RaNDoM6913/claude-code-superkit (2 stars, last pushed 1mo ago), licensed MIT. It adds 39 tokens to every session and 2,366 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 2026-09-03.

Related

Other agents, from other repositories

python-architect

Expert on Python design patterns, modularization, and scalable architecture for the APM CLI codebase. Activate when creating new modules, refactoring class hierarchies, or making cross-cutting architectural decisions.

microsoft/apm · 45 tokens

algorithmic-patterns

Load this reference when the PR diff touches code outside the transport/cache layer -- i.e. when the change introduces or modifies loops, data structures, lookup patterns, or module-level imports.

microsoft/apm · 0 tokens

apm-expert

Expert on APM (Agent Package Manager). Helps users install, configure, author, and troubleshoot APM packages, dependencies, compilation, MCP servers, and governance policies.

microsoft/apm · 39 tokens

test-coverage-expert

Test-coverage expert paired with the DevX UX lens. Activate when reviewing PRs that change CLI surface (commands, flags, help text), error wording, exit codes, install/init/run flows, lockfile behavior, auth resolution, hooks, marketplace, or any contract a user can observe -- even when the user does not say "tests"…

microsoft/apm · 151 tokens

auth-expert

Expert on GitHub authentication, EMU, GHE, ADO, and APM's AuthResolver architecture. Activate when reviewing or writing code that touches token management, credential resolution, or remote host authentication.

microsoft/apm · 46 tokens

oss-growth-hacker

OSS adoption and growth-hacking specialist for microsoft/apm. Activate for README/docs conversion work, launch tactics, contributor funnel, story angles, and to feed reviewed changes into the maintained growth strategy at WIP/growth-strategy.md.

microsoft/apm · 52 tokens