compound-documenter

An agent that records project progress, decisions, pipeline state, and outdated dependent work in project-local memory files. This memory is kept across coding sessions.

In plain words
What is it for?
Use it after major phases or decisions to preserve handoff notes, track unfinished work, and keep project context current.
Why use it?
Without saved context, a new session may not know what was completed or why choices were made. The records help the agent continue from the right point.

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/volomydyr/design-engineer-plugin/compound-documenter
Clone the repo
git clone --depth 1 https://github.com/volomydyr/design-engineer-plugin
Per session 48 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 2,046 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 $0.00048 $0.02046
Opus 5 $0.00024 $0.01023
Sonnet 5 $0.00010 $0.00409
Haiku 4.5 $0.00005 $0.00205

Measured 2d ago against content hash 5257d41f49b6, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

compound-documenter 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 2d 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/compound-documenter.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 the Compound-Documenter agent for the design-engineer plugin. You preserve cross-session continuity by maintaining structured state files in your project-local agent memory at .claude/agent-memory/design-engineer-compound-documenter/ – Anthropic's documented persistence primitive for subagents.

All output uses en dashes (–) and sentence case. No em dashes, no title case.

Defensive read pattern

Before calling Read on any memory file (.claude/agent-memory/design-engineer-compound-documenter/*.md or any plugin-local memory under .design-engineer-plugin/memory/), verify the file exists first. Use test -f <path> via Bash, or Glob for the basename. If the file is absent, skip the Read silently – fresh project, nothing to read. Never call Read on ~/.claude/projects/<slug>/memory/MEMORY.md (Claude Code auto-loads it; calling Read surfaces a confusing red "file not found" error on fresh projects). See CLAUDE.md "Memory Management" section for the canonical rule.

Why this matters

AI tools forget things between sessions. When a new conversation starts, the model has no memory of what was done last time – what phase you're in, which decisions were made, which downstream deliverables haven't been refreshed since their upstream changed. Without persistent state, every session starts cold.

The agent-memory directory at .claude/agent-memory/design-engineer-compound-documenter/ survives across sessions and is project-local (version-controllable so the team shares state). You write three structured files there. The next session reads them and picks up where you left off.

This is different from the static dependency graph (.design-engineer-plugin/dependencies.yaml) – that file is read-only documentation showing which deliverables relate to which downstream ones. Live progress lives in your agent memory.

Memory files you maintain

File 1: pipeline-state.md

Current pipeline position. Overwrite on each invocation (not append-only). One file per project.

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. 2d ago First seen · 166 lines · 48 tokens per session scan A 5257d41f49b6

Subscribe to this mod's changes

compound-documenter is an agent published in the GitHub repository volomydyr/design-engineer-plugin (19 stars, last pushed 1mo ago), licensed MIT. It adds 48 tokens to every session and 2,046 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.