github-communication

A writing guide for GitHub communication, including pull requests, issues, reviews, and replies. It helps explain technical findings in language suited to developers and other readers.

In plain words
What is it for?
Use it before posting a pull request, issue comment, review, review reply, scan summary, or explanation in a GitHub discussion.
Why use it?
It removes unnecessary internal jargon and makes the main finding clear at the start. This helps people understand what changed, what was found, or why no change was made.

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/monarch-initiative/dismech/github-communication
Any agent
npx skills add monarch-initiative/dismech --skill github-communication
Clone the repo
git clone --depth 1 https://github.com/monarch-initiative/dismech

Made for: Claude Code, Codex.

Per session 114 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 669 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 $0.00114 $0.00669
Opus 5 $0.00057 $0.00334
Sonnet 5 $0.00023 $0.00134
Haiku 4.5 $0.00011 $0.00067

Measured 2d ago against content hash 8dbcee45a241, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

github-communication 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.

.claude/skills/github-communication/SKILL.md · 56 lines

How it starts

The opening of the file, as written. The whole thing — 56 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Writing GitHub Comments

The overall goal here is to make GitHub comments in issues and PRs (particularly the first/leading comment) more readable to broader audiences. Avoid too much internal jargon, particularly up-front. Especially avoid elliptical language, use plain terms where possible.

What follows are guidelines, the important thing is the principle. If the issue or PR pertains to a complex technical validation issue, then assume a technical audience, and it will be necessary to use technical language around the schema, LinkML etc. If the issue/PR involves nuanced interpretation of cellular data, then assume a cell/molbio audience. And so on.

Scope. PR bodies, issue comments, review bodies and replies, and the summaries scheduled workflows post. Not in scope: prose inside KB YAML (description, explanation, notes) or in docs/. Those are written for a curator who wants the mechanism, and their denser register is correct — do not apply this skill's plainness rules to them.

The rules

BLUF: Lead with the finding, ideally around 3 sentences. Use plain language, ideally with minimal technical jargon (technical details specific to the intended audience can come later).

Avoid abstract and elliptical language. Avoid metaphors. Be concrete. If it is necessary to use abstractions, ground these with concrete examples.

Say the plain thing when something is wrong. If a decision looks like a mistake, write "I think this is wrong, because X". Do not build an elaborate frame around it. Elliptical prose most often appears where an agent has found a contradiction and is hedging instead of reporting it — the hedge is the defect, not the style.

Use diagrams where appropriate. GitHub will render both mermaid and simple ascii diagrams. Use judgment. With mermaid you have less control of layout, and it will be less compact, but it can be easier to grasp on a first pass.

No meta-commentary. Do not narrate your process, list what you could not check unless it changes the conclusion, apologize for length, or tell the reader you tried to write plainly. Saying you were brief is not being brief.

Read the full file on GitHub · 56 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 · 56 lines · 114 tokens per session scan A 8dbcee45a241

Subscribe to this mod's changes

github-communication is a skill published in the GitHub repository monarch-initiative/dismech (56 stars, last pushed 2d ago), licensed BSD-3-Clause. It adds 114 tokens to every session and 669 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 2026-08-30.

Related

Other skills, from other repositories

boss

Use ONLY when the user explicitly asks to orchestrate parallel external agents (Codex or Claude Code) in tmux sessions via tp (tmux-pilot). Or when a user tells you that you're the boss or orchestrator. You can also use this if you are an operclaw agent. NEVER auto-invoke for a generic "do this in parallel" request.…

ai4curation/ai-gene-review · 95 tokens

rule-reviewer

Use this agent when you need to review UniProt annotation rules (ARBA, UniRule) for quality, biological accuracy, and GO annotation appropriateness. This agent performs comprehensive analysis of rule condition sets, evaluates literature support, assesses taxonomic scope, and recommends curation actions.

ai4curation/ai-gene-review · 60 tokens

bioreason-predictions

Work with BioReason-Pro predictions and reasoning traces for protein function. Use when: comparing BioReason outputs against curated reviews, converting web exports to YAML, reviewing reasoning trace quality, or running the SFT/RL comparison. Triggers: "bioreason", "GO-GPT predictions", "reasoning trace review"…

ai4curation/ai-gene-review · 83 tokens

pathway-inference-agent

Use this agent when you need to create a pathway summary for a gene after completing the main gene review. This agent should be invoked specifically to generate GENE-pathway.md files that summarize the gene's role in biological pathways with proper citations and mermaid diagrams.

ai4curation/ai-gene-review · 58 tokens

review-function-prediction

Review computational protein or gene function predictions and fill PredictionReview YAML files using the COR/CNN/LSP/UNC/PLI/NPI/REP biological-validity taxonomy. Use for DeepECTF, BioReason/GO-GPT, InterPro2GO, PANTHER/IBA, CLEAN, GloEC, MAPred, ProteinInfer, or other predicted EC/GO annotations.

ai4curation/ai-gene-review · 83 tokens

core-function-synthesizer

Use this agent when you need to synthesize and distill the core functions of a gene from multiple information sources (textual summaries, reviewed annotations, bioinformatics analysis) and populate the corefunctions section of a gene review YAML file. This agent should be called after completing the gene research…

ai4curation/ai-gene-review · 81 tokens