Skip to main content
Monocle ships standardized SKILL.md files in the agentskills.io format. These files are embedded in the Monocle binary and installed into agent-specific directories when you run monocle register. Skills instruct agents to call monocle review ... CLI commands at the right times.

​
Skills overview

SkillAvailable forDescription
/get-feedbackClaude Code, OpenCode, Codex CLI, Gemini CLIRetrieve pending review feedback from the queue
/get-feedback-waitClaude Code, OpenCode, Codex CLI, Gemini CLIBlock until the reviewer submits feedback, then act on it
/review-planClaude Code, OpenCode, Codex CLI, Gemini CLIFind the active plan file and submit it for review — returns immediately
/review-plan-waitClaude Code, OpenCode, Codex CLI, Gemini CLIFind the active plan file, submit for review, and iterate until approved

​
Where skills are installed

monocle register installs skills into the agent’s skills directory for the current project. Use --global to install into the user-level config directory instead.
AgentProject skills directoryGlobal skills directory
Claude Code.claude/skills/~/.claude/skills/
OpenCode.opencode/skills/~/.config/opencode/skills/
Codex CLI.codex/skills/~/.codex/skills/
Gemini CLI.gemini/skills/~/.gemini/skills/

​
Skill details

​
/get-feedback

Retrieves pending review feedback from Monocle. Use this after the reviewer has submitted feedback through the TUI, or after receiving a feedback_submitted push notification. Underlying command: 
monocle review get-feedback
Agent behavior:
  • If feedback is available, the agent reads it carefully and acts on the reviewer’s comments, issues, and suggestions.
  • If no feedback is pending, the agent informs you that no review is available yet.
  • If the command fails because Monocle is not running, the agent tells you to start Monocle with monocle in the project directory.

​
/get-feedback-wait

Blocks until the reviewer submits feedback through Monocle, then acts on it. Use when a pause has been requested, or when the agent should wait for reviewer approval before continuing. Underlying command: 
monocle review get-feedback --wait
Agent behavior:
  • The agent blocks on the command until you submit a review.
  • After receiving feedback, the agent addresses the comments in your code.
  • If you requested changes, the agent runs monocle review get-feedback --wait again after making updates, and keeps iterating until you approve.

​
/review-plan

Finds the active plan file in the project and submits it to Monocle for the reviewer to see. Returns immediately — does not wait for feedback. Use /review-plan-wait if you need the agent to block until you respond.
This skill is for content that isn’t already a tracked file change — plans, architecture docs, summaries, etc. Monocle automatically picks up code file changes in the workspace. You do not need to send those.
Underlying command: 
monocle review send-artifact \
  --title "<first heading or filename>" \
  --file <absolute path to plan> \
  --id <plan filename> \
  --type md
Agent steps:
  1. Find the plan file — uses the path from $ARGUMENTS if provided, otherwise finds the most recently modified plan file in the project.
  2. Read the file to confirm it exists and extract its filename.
  3. Run monocle review send-artifact with the title set to the first markdown heading (or the filename if no heading is found). The --id is set to the plan filename so that re-submissions replace the previous version rather than creating duplicates.
  4. Confirm to you that the plan was sent to Monocle.

​
/review-plan-wait

Finds the active plan file, submits it to Monocle, and blocks until the reviewer responds with feedback. If you request changes, the agent updates the plan and resubmits — iterating across as many rounds as it takes until you approve. Underlying command: 
monocle review send-artifact \
  --title "<first heading or filename>" \
  --file <absolute path to plan> \
  --id <plan filename> \
  --type md \
  --wait
Agent steps:
  1. Find the plan file (same logic as /review-plan).
  2. Read the file to confirm it exists.
  3. Run monocle review send-artifact --wait — this submits the artifact and then blocks until you submit a review.
  4. Handle the response:
    • If you approved with no comments, the agent informs you and continues.
    • If you provided feedback requesting changes, the agent shares the feedback with you, updates the plan, and re-runs monocle review send-artifact --wait.
    • The agent keeps iterating until you approve.

​
How skills work

Each skill is a SKILL.md file that describes when and how to call monocle review CLI commands. When an agent encounters a skill invocation (e.g. /review-plan), it reads the skill file and follows the steps — which ultimately call monocle review ... to communicate with the running Monocle TUI over a Unix socket. Skills are wrappers for convenience. Agents can also call monocle review ... commands directly without using a skill — for example, adding monocle review get-feedback to a step in a custom agent prompt.

​
Triggering skills manually

You can invoke /review-plan and /review-plan-wait yourself at any time by typing the skill name in your agent’s chat interface. This is useful when you want to send the agent’s current plan for review without waiting for the agent to do it on its own.
To make the agent automatically submit plans for review, add instructions to your project’s agent config file (e.g. CLAUDE.md, AGENTS.md). See Agent integration for a suggested prompt.

​
Skills vs direct CLI

SkillsDirect CLI
How invokedAgent reads SKILL.md and follows stepsAgent calls monocle review ... in a shell step
Setup requiredmonocle registerNone — any agent with shell access works
Best forAgents with skill supportCustom prompts, scripts, or agents without skill systems
Both approaches call the same underlying monocle review ... commands and connect to the same Unix socket.

CLI reference

Full reference for all monocle review commands used by skills.

Agent integration

How to register Monocle with your agent and configure automatic reviews.

Build docs developers (and LLMs) love

Get started for freeTalk to us