adr-writer

adr-writer is an agent for Claude Code from shimo4228/claude-harness. It costs 124 tokens per session (2,067 once invoked), scanned A, original, MIT.

A document writer for Architecture Decision Records, or ADRs—short files that record an important software design choice and why it was made.

In plain words
What is it for?
Use it to render one approved decision packet into an ADR with sections for status, date, context, decision, review conditions, alternatives, and consequences.
Why use it?
It turns an already-decided design decision into a consistent record without making up missing reasoning. This leaves future developers a clear account of the decision and its trade-offs.

Agent for Claude Code

Written for Claude Code: a Claude Code subagent (agents/*.md). Also seen: model in frontmatter; reads .claude/ paths.

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/shimo4228/claude-harness/adr-writer
Clone the repo
git clone --depth 1 https://github.com/shimo4228/claude-harness

Made for: Claude Code.

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 adr-writer

README.md
[![agentmods](https://agentmods.dev/badge/agents/shimo4228/claude-harness/adr-writer.svg)](https://agentmods.dev/agents/shimo4228/claude-harness/adr-writer)
Your own site
<a href="https://agentmods.dev/agents/shimo4228/claude-harness/adr-writer"><img src="https://agentmods.dev/badge/agents/shimo4228/claude-harness/adr-writer.svg" alt="Measured on agentmods" height="20"></a>
Per session 124 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 2,067 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 5.1 $0.00124 $0.02067
Opus 5 $0.00062 $0.01033
Sonnet 5 $0.00025 $0.00413
Haiku 4.5 $0.00012 $0.00207

Measured 6d ago against content hash 7cde51cc85dd, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

adr-writer 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 6d 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/adr-writer.md · 166 lines

How it starts

The opening of the file, as written. The whole thing — 166 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are an ADR writer. Your job is to take a developer's raw description of a design decision and produce a single ADR file with the 7 canonical sections filled in, matching the style and rigor of existing ADRs in the target repo.

Core Principle

Refactor the user's words. Never invent.

If the caller did not give you a Context, a Decision, a Review-when, an Alternatives Considered, or a Consequences, stop and ask. Padding sections with plausible-sounding fabrication is the failure mode this agent exists to prevent.

You hold no semantic authority. The decision packet (Context / Decision / Review-when / Alternatives + rejection reasons or revisit conditions / Consequences) is decided and approved by the caller — the main loop — before you are invoked. Your job is rendering that frozen packet into the template: re-expressing, reorganizing, matching house style, resolving links, writing the file. You do not decide what the ADR means, and you do not infer content the caller did not supply. See ADR-0016 — writer agents render, they do not decide.

Input You Will Receive

The caller (typically adr-writer skill) passes:

  • ADR number (already resolved, e.g., 0010)
  • Repo root (absolute path)
  • ADR directory (e.g., <repo-root>/docs/adr/)
  • Title (kebab-case, e.g., context-sync-cascade-and-writer-agents)
  • Status (one of: proposed | accepted | superseded | deprecated; default accepted)
  • Date (ISO 8601, default today)
  • Context (raw text — what problem prompted the decision)
  • Decision (raw text — what was decided)
  • Review-when (raw text — the expiry conditions: which observation or premise failure would void or weaken this decision; or the explicit statement that there is none)
  • Alternatives (raw text or list — what else was considered and why rejected, or, for an alternative that stays live, the condition under which it would be revisited)
  • Consequences (raw text — what becomes easier / harder)

Read the full file on GitHub · 166 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. 6d ago First seen · 166 lines · 124 tokens per session scan A 7cde51cc85dd

Subscribe to this mod's changes

adr-writer is an agent published in the GitHub repository shimo4228/claude-harness (2 stars, last pushed 3d ago), licensed MIT. It adds 124 tokens to every session and 2,067 once invoked, about $0.0006 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

code-documentation-docs-architect

Creates comprehensive technical documentation from existing codebases. Analyzes architecture, design patterns, and implementation details to produce long-form technical manuals and ebooks. Use PROACTIVELY for system documentation, architecture guides, or technical deep-dives.

wshobson/agents · 55 tokens

whitepaper-coherence

Analyse la cohérence globale d'un livre blanc (logique, contradictions, ruptures narratives, redondances). Utiliser pour auditer un whitepaper avant publication.

FlorianBruniaux/claude-code-ultimate-guide · 39 tokens

mb-reviewer

Code review agent for /mb work review-loop. Reads stage diff + pipeline.yaml reviewrubric and emits structured JSON verdict (APPROVED / CHANGESREQUESTED) with severity-classified issue list. Drives the severity-gate decision.

fockus/skill-memory-bank · 53 tokens

mb-judge

Independent final quality gate for Memory Bank governed workflows. Decides GO, GOWITHBACKLOG, or NOGO after verifier and lead-review reports.

fockus/skill-memory-bank · 33 tokens

mb-architect

Architecture / ADR / system-design specialist for memory-bank /mb work stages. Domain modelling, interface definition, ADR authoring, refactoring strategy. Does not ship features alone.

fockus/skill-memory-bank · 39 tokens

mb-qa

QA / testing specialist for memory-bank /mb work stages. Test design, coverage strategy, edge-case enumeration, flake elimination, contract tests. Falls back to mb-developer when stage is generic.

fockus/skill-memory-bank · 44 tokens