[FEATURE] Request: Option to hide/disable full skill content display in conversation

Description

Currently, when an agent loads a skill using the skill tool, the full content of the SKILL.md file is displayed in the conversation. While the tool description only shows the skill name and description (a brief version), the complete skill content takes up significant screen space once loaded.

Problem

This behavior can be disruptive to the conversation flow, especially when:

The skill content is lengthy
Multiple skills are loaded during a session
Users want to focus on the actual conversation without scrolling through large skill content blocks

Requested Feature

Please consider adding a configuration option to control how skill content is displayed. Some potential solutions:

Collapse skill content by default - Add a toggle to show skill content in a collapsible/expandable format
Show only skill name - Option to display just the skill name when it's loaded, without the full content
Disable skill content display - A setting to completely hide skill content from the conversation view (while still making it available to the agent)
Compact mode - Show skill content in a more compact format

Expected Behavior

The skill tool already shows a brief version (<available_skills>) in the tool description. It would be great if users could control whether the full content is displayed in the conversation after the skill is loaded.

Environment

OpenCode version: [latest]
Terminal: [your terminal]

References

Skill tool documentation: https://opencode.ai/docs/skills/
Current behavior shows full SKILL.md content in conversation after skill({ name: \"skill-name\" }) is called

This would be a quality-of-life improvement for users who frequently use skills and want to keep their conversation view cleaner.

I have searched existing feature requests and confirmed this is not a duplicate

This issue doesn't fully meet our contributing guidelines.

What needs to be fixed:

Feature requests must use the Feature Request issue template, with a title starting with [FEATURE]: and include the required verification checkbox confirming no duplicate exists

Please edit this issue to address the above within 2 hours, or it will be automatically closed.

This issue might be a duplicate of existing issues. Please check:

配置选项：隐藏 skill 工具调用返回的内容显示 #15427: Same request — a configuration option to hide/collapse full skill content displayed in the conversation after the skill tool loads a skill

If you believe this was flagged incorrectly, please let a maintainer know.

Thanks for updating the issue! It now meets the contributing guidelines. The feature request is clearly described with a well-defined problem and proposed solutions.

Thanks for updating the issue! It now meets our contributing guidelines. The feature request has a clear title, the required verification checkbox, and a well-described problem with proposed solutions.

I think this would be a cool feature for not just skills, but any chat prompt / response! If I paste in ~200lines and I want to quickly review the chat prior to my long message, there's a lot of scrolling involved. Same applies to agent outputs

@thdxr

@thdxr When can I wait for an answer to this question?

@thdxr

I'd like to take a crack at this. The fix is minimal - in resolvePromptParts() in packages/opencode/src/session/prompt.ts, the template text part can be marked synthetic: true:

const parts: PromptInput["parts"] = [
  {
    type: "text",
    text: template,
    synthetic: true,
  },
]

The synthetic flag already has the right semantics - it's used throughout the codebase to mark content that's sent to the model but hidden from the user (file reads, bookkeeping messages, etc.). The filtering is already in place everywhere parts are rendered, so no UI changes are needed in either the TUI or the web/desktop app. I've validated this locally in OpenWork as well and it hides the command=>skill prompt parts as desired.

Optionally, a companion ignored: true part could show a short confirmation like "Running skill: skill-name" so the user gets visual feedback without seeing the full template. I've validated this locally in OpenWork as well - as an example ->

Happy to open a PR if this approach works for you. cc @thdxr

Navigation Menu

Repository navigation

[FEATURE] Request: Option to hide/disable full skill content display in conversation #15805

Description

Problem

Requested Feature

Expected Behavior

Environment

References

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

[FEATURE] Request: Option to hide/disable full skill content display in conversation #15805

Description

Description

Problem

Requested Feature

Expected Behavior

Environment

References

Activity

github-actions commented on Mar 3, 2026

github-actions commented on Mar 3, 2026

github-actions commented on Mar 3, 2026

ericmiranda7 commented on Mar 10, 2026

UoMaster commented on Mar 20, 2026

bdruth commented on Mar 29, 2026

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions