plannotator-last

Collects user annotations on the latest assistant message via the Plannotator CLI.

Updated Aug 22, 2026
One-click install
npx skills add https://github.com/farslab/claude-skills --skill plannotator-last-farslab
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: plannotator-last
Source: https://github.com/farslab/claude-skills/tree/main/plannotator-last
Command: npx skills add https://github.com/farslab/claude-skills --skill plannotator-last-farslab

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Reviewing an AI assistant's latest message often requires back-and-forth clarification. This Skill opens Plannotator on the most recent rendered assistant message so the user can approve it, dismiss it, or attach inline annotations that guide the next revision. ## Core Features & Use Cases - Message Annotation: Runs plannotator annotate-last to open an annotation session on the latest assistant message. - Decision Handling: Interprets three outcomes—approval (optionally with non-blocking feedback notes), dismissal, or annotated feedback—and responds accordingly. - Use Case: After the assistant drafts a plan or proposal, invoke this Skill so the user can mark up specific sections; the assistant then revises the message based on the returned annotations. ## Quick Start Ask the assistant to open Plannotator on its last message so you can review and annotate it.

Frequently Asked Questions about plannotator-last

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I annotate the last assistant message with Plannotator?▼

Invoke the Skill, which runs plannotator annotate-last to open an annotation session on the most recent rendered assistant message. You can then approve it, dismiss the session, or leave inline comments that the assistant uses to revise the message.

What happens after I approve a message in Plannotator?▼

Approval returns either the text "The user approved." or a JSON object with decision "approved". The assistant acknowledges with a single sentence and stops; if the approval includes a feedback field, those notes are carried forward as non-blocking guidance.

Does plannotator-last require any external tools?▼

Yes, it requires the plannotator CLI to be available in the environment, since the Skill executes plannotator annotate-last through Bash. No other scripts or package dependencies are included in the Skill directory.

What does it mean when the annotation session returns empty output?▼

Empty output or a JSON decision of "dismissed" means the user closed the session without requesting changes. The assistant acknowledges with a single sentence and does not begin any work.

When should I use annotated feedback instead of approval notes?▼

Use annotated feedback when you want the assistant to revise its last message based on specific comments. Approval notes are non-blocking guidance for future work, while annotations request concrete changes to the reviewed message.