self-communicate

A writing and diagramming guide for technical material read by engineers and coding agents. It covers document types, structure, vocabulary, tone, and reviews of prose and diagrams.

In plain words
What is it for?
Use it to write or review design documents, runbooks, handoffs, READMEs, tutorials, architecture decisions, API references, control descriptions, and status reports.
Why use it?
It helps make technical instructions, decisions, and explanations easier to follow and more consistent. It also helps match a document to its purpose, such as teaching, reference, or operational guidance.

Skill for Claude CodeCodex

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 skills/davisjam/model-based-agentic-software-engineering/self-communicate
Any agent
npx skills add davisjam/model-based-agentic-software-engineering --skill self-communicate
Clone the repo
git clone --depth 1 https://github.com/davisjam/model-based-agentic-software-engineering

Made for: Claude Code, Codex.

Per session 286 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 4,224 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 $0.00286 $0.04224
Opus 5 $0.00143 $0.02112
Sonnet 5 $0.00057 $0.00845
Haiku 4.5 $0.00029 $0.00422

Measured yesterday against content hash 7d41afc7e302, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

self-communicate 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.

The scan reads SKILL.md. This mod also ships 2 executable files (drawing/svg-audit.py, manifest.py), listed below but not scanned — reading those needs a real analyzer, not pattern matching.

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.

plugin/mage/skills/self-communicate/SKILL.md · 234 lines

How it starts

The opening of the file, as written. The whole thing — 234 lines — stays where its author put it; the contents beside it link to each section on GitHub.

self-communicate — write, draw, and present engineer-facing material well

You are communicating the way the work is explained in this repository — the prose and diagrams an engineer or a coding agent reads to understand it. This is the communicate leg of a trio: govern / operate / communicate. self-governance is the harden leg (the census of controls and the engine that mints new ones); self-operations is the operate leg (running the substrate those controls govern). This skill governs the prose the fleet and its operator produce: the docs those two write — control descriptions, design docs, mechanism entries, runbooks, handoffs — and the operator's own communication with the human — a status report, a tradeoff you explain, a decision you surface. The orchestrator's prose to the user is prose too. Make all of it clear, consistent, and readable.

The organizing idea: good technical prose is a craft with named parts, not a matter of taste. Rhetoric supplies the devices; Diátaxis fixes the doc's shape; a lexicon fixes the vocabulary; a voice fixes the register; an audit procedure grades the result. Diagrams are the same craft for shapes prose can't carry. None of this is invented per document — you reach into a toolkit and apply it.

The governing stance: less is more — the representation must not distract from the idea

One principle governs every resource below — across all three legs: the representation is a carrier for the idea; strip anything that does not carry it. Three touchstones name the discipline:

  • Hemingway — economy of words. Plain, concrete, load-bearing prose. No fluffy adjective, no ornamental flourish where a plain word carries the idea.
  • Tufte — the data-ink ratio. A visual aid is the simplest form that carries its idea; strip the chartjunk — the gridlines, gradients, and 3-D bevels that decorate without informing.
  • Early Picasso, the bull lithographs — the eleven-plate Bull reduces a bull to a few lines, each plate removing what the last could spare, until only the essence remains. That reduction is the target for a diagram: the fewest marks that still read as the thing.

Read the full file on GitHub · 234 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 · 234 lines · 286 tokens per session scan A 7d41afc7e302

Subscribe to this mod's changes

self-communicate is a skill published in the GitHub repository davisjam/model-based-agentic-software-engineering (18 stars, last pushed yesterday), licensed MIT. It adds 286 tokens to every session and 4,224 once invoked, about $0.0014 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.