Release Comms logo

Release Comms

CommunityPopular
huytieu
release-comms

Release communication pipeline for your product: turn a deployment runsheet and a GitHub milestone into a customer-facing release note with real screenshots and videos, docs updates (including retrofits of articles the release makes stale), a Slack announcement draft, and a lightweight marketing package. Use when you say "cover today's release", "write the release note", or "/release-comms".

Overview

Publisherhuytieu
RepositoryCOG-second-brain
Skill namerelease-comms
Stars
1.3K
Forks
162
Bundled files
Instructions only
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by huytieu on GitHub. Read the source before you install it.

Installation

Install the Release Comms AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/huytieu/COG-second-brain/tree/main/.claude/skills/release-comms
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/huytieu/COG-second-brain.git /tmp/COG-second-brain
mkdir -p .claude/skills
cp -r /tmp/COG-second-brain/.claude/skills/release-comms .claude/skills/release-comms
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Release Comms in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Release Comms on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Release Comms is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

release-comms: release note, docs, announcement and marketing pipeline

Lane: full. It mutates external state (Slack posts, docs PR, wiki pages), so every post and PR is drafted first and gated on the user's explicit approval. See the closed-loop skill.

Configuration

Fill these in once (kit.toml, .env or arguments). Placeholders below are examples.

SettingMeaningExample
PRODUCT_REPORepo that holds the release milestoneyour-org/your-product
MILESTONE_PATTERNMilestone namingRelease YYYY-MM-DD
TEAM_LABELLabel that filters the milestone to your teamteam:platform
RUNSHEET_SOURCEWhere the deployment runsheet livesa wiki page, a doc, or a Slack thread
RELEASE_DIRWorking folder per releasecomms/release-notes/<date>/
DOCS_REPOLocal path of the docs repo../your-docs
RELEASE_NOTES_FILERunning release-notes page in the docs repodocs/release-notes.md
ANNOUNCE_CHANNELSlack channel for the internal announcement#product-releases
MARKETING_CHANNELSlack channel for the marketing package#marketing
FLAG_SYSTEMFeature-flag tool named in the runsheetany

Source-of-truth chain (walk it in order)

  1. Deployment runsheet. Gives release scope, the milestone link, the PR list per service, feature flags being enabled, the deploy window, and trailing items (merged to QA but not staged). Exclude trailing items from "available now".
  2. GitHub milestone. The milestone filtered by TEAM_LABEL is the release manifest. Issues may still be open, so closed does not equal shipped.
  3. Issue bodies and comments. This is the media source. Dev-test comments ("Dev test: passed on prod") and QA comments ("Verified PASSED on staging") carry screenshots and demo videos as GitHub attachment URLs. For a private repo these need auth to resolve: curl -sL -H "Authorization: Bearer $(gh auth token)" -o <file> <url> Check the Content-Type with curl -sI for the right extension. GitHub videos are usually QuickTime .mov. Treat a comment-level "Verified PASSED" as the ship signal, because issues often stay open until a later close step.
  4. Feature flags. Anything behind a flag in the runsheet gets honest gating language ("rolling out", "enabled per organization").

Pipeline

1. Collect (parallel workers, staging files)

Use one worker per headline cluster plus one sweep worker over the remaining milestone issues. Each worker extracts user-visible changes with verbatim quotes and links, classifies each as user-facing feature, user-facing fix or internal, downloads media into assets/ with descriptive kebab-case names, writes to RELEASE_DIR/staging/<cluster>.md, and returns only a status line. Never paste one worker's output into another's prompt.

2. Release note (lead writes, single file)

Write RELEASE_DIR/release-note-<date>.md: frontmatter (version, date, milestones, labels), a one-paragraph narrative summary, Highlights (bold, benefit-first bullets with issue links), per-feature sections (What's new, Why this matters, What you can do now, Shipped in this capability), Enhancements, Fixes, and an appendix table of shipped issues. See the generate-release-notes skill for the base format.

Voice rules:

  • Do not write a feature list. Every item says what the user can now do and what pain it removes. Lead with the before/after story where one exists.
  • Embed the downloaded screenshots and videos inline next to the feature they prove.
  • Disclose known issues plainly with a fix timeline.
  • Flag to the user anything user-benefiting found in the sweep that they did not list.
  • Frame the release around the opportunity it opens, and sanity-check every number.
  • Naming: avoid "v1" or "first version" in customer copy. Call gated first releases "early access". Call UI refreshes "a new look", not "redesign". Describe capabilities as what is now supported, and say when one serves both humans and agents.
  • The internal post is the deep dive: same naming, plus issue links, flag names, architecture notes, verification status and known-issue numbers. Marketing copy stays light.

3. Docs updates (branch off main, never touch in-flight branches)

Use a git worktree so work in the main checkout is never disturbed: git worktree add ../docs-release-<date> -b release-comms/<date> origin/main (run inside DOCS_REPO). Check what exists on origin/main before planning edits, because drafts in the checkout may not be merged. Two motions:

  • New or changed feature docs. Use the feature-doc skill for any new article or real screenshot refresh.
  • Retrofit sweep. Grep the docs for surfaces this release changed: stale screenshots, stale flow text, stale limitations. Always add the release entry to RELEASE_NOTES_FILE (heading per date, "What's new" and "Why this matters" for features, enhancements and fixes in collapsible blocks). If unmerged drafts cover the same surface, do not edit them. Flag the needed change to the user for that branch. The docs PR is a ship gate: the user approves before it opens. To publish the note to a wiki, see the publish-to-confluence skill.

4. Slack announcement draft (ANNOUNCE_CHANNEL)

Format: an announcement line (:mega: *Product Release Announcement* | <release note link> | Available now), a channel-wide greeting, then emoji-prefixed bold-name bullets with one benefit sentence each, a known-issue heads-up, and a link to the full release note. Keep it terse. Draft only: the user posts or approves posting.

5. Marketing package (MARKETING_CHANNEL)

Use a lightweight changelog-thread format, not positioning essays:

  1. Lead post: "This week's release:" plus a numbered list (1/ ...), one line per shipped feature, and one short closing line. No media required.
  2. One detail post per feature: a hook line (the pain or the punch), 2 to 4 short lines on what it does, the benefit and a use case, an honest caveat if any, and a specific demo video, GIF or screenshot from assets/ (name the file and note any trim or convert work). Each post reuses cleanly as an X or LinkedIn thread, a Slack lead with thread replies, or a newsletter section.
  3. Media manifest: each asset marked near-ready, needs trim, or missing (record fresh). The release-video, social-media-kit and youtube-video-kit skills cover producing media.
  4. Suggested channel message for the people who compile and review the newsletter. Draft only: gate on the user.

5b. Demo videos back to the tickets (optional)

Once demo videos are published somewhere the team can open, write each video link into a "Demo" field on every shipped issue it covers (for example a GitHub Project text field, set through gh api graphql with updateProjectV2ItemFieldValue). One link per issue. Check an existing value before overwriting it. Verify by re-reading the field.

6. Verify (closed loop)

  • A fresh-context verifier re-checks release-note claims against the issue links and tags each Verified, Weakened or Falsified. Drop or demote accordingly.
  • Media check: every embedded asset exists locally and renders (size above zero, correct extension).
  • After any approved Slack post or docs PR, re-fetch the artifact and confirm it matches intent.

Cockpit

Keep a session-review.md in RELEASE_DIR (the review-cockpit pattern): progress, one card per deliverable, the user's approvals logged. Delete staging files after consolidation.

Frequently asked questions

What does the Release Comms AI skill do?

Release communication pipeline for your product: turn a deployment runsheet and a GitHub milestone into a customer-facing release note with real screenshots and videos, docs updates (including retrofits of articles the release makes stale), a Slack announcement draft, and a lightweight marketing package. Use when you say "cover today's release", "write the release note", or "/release-comms".

Why use Release Comms on TypingMind?

Because you install it once and use it with any model. Release Comms is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Release Comms in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/huytieu/COG-second-brain/tree/main/.claude/skills/release-comms. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Release Comms?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Release Comms?

As many as you like. As long as a model supports skills, you can use Release Comms with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Release Comms AI skill free?

Yes. It is published on GitHub by huytieu under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇