Revyl Cli Atlas Review logo

Revyl Cli Atlas Review

Organization
RevylAI
revyl-cli-atlas-review

Inspect exact Atlas evidence and manage grounded annotation feedback when the user explicitly requests a feedback mutation.

Overview

PublisherRevylAI
Repositoryrevyl-cli
Skill namerevyl-cli-atlas-review
Stars
522
Forks
25
Bundled files
1
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.

  • 1 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

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

Installation

Install the Revyl Cli Atlas Review 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/RevylAI/revyl-cli/tree/main/skills/revyl-cli-atlas-review
  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/RevylAI/revyl-cli.git /tmp/revyl-cli
mkdir -p .claude/skills
cp -r /tmp/revyl-cli/skills/revyl-cli-atlas-review .claude/skills/revyl-cli-atlas-review
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Revyl Cli Atlas Review 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 Revyl Cli Atlas Review 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 Revyl Cli Atlas Review 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.

Revyl Atlas Review Skill

Use this write-capable leaf only when the user explicitly asks to add or manage Atlas feedback. Inspect first with revyl-cli-atlas: resolve the app, open the relevant screenshots, select the exact observation, and list existing feedback before mutating anything.

For each reviewed screen, expand variants with --include-variants on screen and observations inspection, and open the relevant state screenshots before forming a finding. A representative image does not stand for every variant. Anchor feedback to the exact inspected observation, including when the issue only occurs in a variant; never generalize it to uninspected states.

Treat the screenshot and pin as baseline context. Attach supporting media or files only when they add information the anchored view cannot provide—such as a transition, comparison, log, report, or implementation note—and optimize for decision value rather than volume. Inspect and use existing attachments as first-class context before replying, editing, or resolving; do not attach a duplicate screenshot that merely repeats the pinned view.

Write for scanning

Put one finding in each annotation. Start with <Type>: <short summary> that stands on its own, ideally 3–8 words after the colon. Use Blocker, Issue, or Polish for supported problems; Feedback for a suggestion and Question for uncertain intent. These are body labels, not new severity values.

Follow with at most two short sentences only when they add necessary context: what differs, the user impact, or the decision needed. Aim for 40 words or fewer in total; a clear headline alone is enough. Cut preambles, repeated screen descriptions, IDs, test narration, and speculative explanations. Preserve any uncertainty that changes the conclusion. Do not use em dashes.

text
Issue: Save button does nothing
Tapping Save leaves the form unchanged. Expected the updated profile.

Question: Card title wraps differently
The reference uses two lines; this state uses three. Is that intentional?

The pin provides the app screenshot. Add a relevant design crop, comparison, or clip when it helps explain the finding; put longer analysis in an attachment or linked report instead of the comment. Do not duplicate the pinned screenshot. Replies lead with the new result or decision, such as Still present: Save has no effect, and add only new context rather than repeating the root finding.

Ground the annotation

Use concrete visual targets such as “the blue Continue button at the bottom,” not semantic guesses. Preview ambiguous targets before creation or movement:

bash
PREVIEW_FILE=$(mktemp -t atlas-annotation-preview.XXXXXX.png)
revyl atlas annotations create \
  --app <app-id> \
  --observation <observation-id> \
  --target "<visible element and location>" \
  --dry-run \
  --preview-out "$PREVIEW_FILE" \
  --json

Open the marked preview and verify the pin. Create only after it is correct:

bash
revyl atlas annotations list --app <app-id> --observation <observation-id> --status all --json
revyl atlas annotations create --app <app-id> --observation <observation-id> \
  --target "<visible element and location>" --body "<actionable feedback>" \
  --severity <blocker|issue|polish> --attach <evidence-path> \
  --client-request-id <saved-uuid> --json

When the annotation reports a problem, set --severity: blocker blocks shipping, issue is wrong but shippable, polish is cosmetic. Omit severity for questions and discussion; change it later with revyl atlas annotations severity <thread-id> --app <app-id> --severity <value> (or --clear).

Generate and save a UUID before each create/reply and pass it through --client-request-id on the first attempt. Capture stdout, stderr, and exit status separately; parse stdout as JSON without 2>&1. The request-ID diagnostic belongs to stderr. Empty or unparseable output is not proof the write failed: inspect the captured results and read back server state before retrying. If a retry is still necessary, use the same saved UUID, body, target, observation, and ordered attachment paths. Changing the payload with that ID is a conflict. Never reissue a mutation with a new UUID merely to debug its output. Do not automatically delete suspected duplicates; reconcile and report the state first. Repeat --attach for up to four files. Edit uses --attach, repeatable --remove-attachment, and --clear-attachments; combining clear and attach replaces the attachment set, while omitting attachment flags preserves it. To mention a human organization member, discover their ID first, then bind a local alias to one placeholder in the body:

bash
revyl atlas annotations members --app <app-id> --query <name-or-email> --json
revyl atlas annotations reply <thread-id> --app <app-id> \
  --body '@{reviewer} can you check this state?' \
  --mention 'reviewer=<user-id>' --client-request-id <saved-uuid> --json

Mention the smallest set of relevant stakeholders whose ownership, expertise, approval, or action is needed. Good candidates include the owner of the affected product or code surface, a designer or engineer needed to answer a specific question, and an existing thread participant needed to make a decision. Do not mention every organization member for visibility alone. Give each mentioned person a concrete reason to engage, and omit the mention when the comment is informational and requires no response.

Use the same @{alias} and repeatable --mention alias=user-id syntax with create, reply, and edit, including body-file or stdin input. Bind each alias and member once; unresolved placeholder-like text remains literal. The CLI replaces bound placeholders with the member's current display name and submits structured mention spans. Only human organization members can be mentioned. Never automatically retry a version conflict: read the current thread and decide against that state. Move always grounds against the thread's immutable observation.

Use list one page at a time and follow next_cursor deliberately. Statuses are open, resolved, or dismissed; closed aggregates the latter two. Deleting always requires --yes. Before deleting, remember that deleting a root comment removes the full thread from internal and public-share surfaces.

Return the focused Atlas URL after a successful mutation. Keep customer screenshots and marked previews temporary and never expose signed URLs, bodies, targets, or request IDs in committed artifacts.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Revyl Cli Atlas Review AI skill do?

Inspect exact Atlas evidence and manage grounded annotation feedback when the user explicitly requests a feedback mutation.

Why use Revyl Cli Atlas Review on TypingMind?

Because you install it once and use it with any model. Revyl Cli Atlas Review 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 Revyl Cli Atlas Review in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/RevylAI/revyl-cli/tree/main/skills/revyl-cli-atlas-review. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Revyl Cli Atlas Review?

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 Revyl Cli Atlas Review?

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

Is the Revyl Cli Atlas Review AI skill free?

It is published on GitHub by RevylAI. Check the repository for licensing terms. 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 👇