Command Management Skill

SkillDev tools

Use when a user wants to create, update, list, enable, disable, or delete Space or project slash commands through natural language.

Available today. Use it from your connected AI after setup.

Connect ahel once, and every AI you use reads what you have installed.

Then ask your AI: use the Command Management Skill skill

What this skill tells your AI

The instructions your AI receives, as published by contextgo/contextgo in src/process/resources/skills/_builtin/command-management/SKILL.md and read by ahel’s review.

You can manage ContextGo commands for either the current project or the bound Space.

Rules

  1. Ask which scope to use if the user did not make it clear: project or space.
  2. Use project for workspace-private commands and overrides in .contextgo/commands.json.
  3. Use space for shared commands stored with the bound Space.
  4. Update by slash name within the selected scope. If /review already exists in that scope, overwrite it directly.
  5. Delete by slash name within the selected scope.
  6. Ask for confirmation in the conversation before destructive deletion or broad overwrites when the user intent is ambiguous.
  7. Output control commands directly. Do not wrap them in markdown code fences.

Command Formats

List commands

Output:

[COMMAND_LIST: scope=project]

or

[COMMAND_LIST: scope=space]

Create or update a command

Output:

[COMMAND_UPSERT] scope: project name: review enabled: true description: Review the current diff for regressions and missing tests. template: Review the current changes like a strict code reviewer. Prioritize bugs, regressions, risky assumptions, and missing tests. [/COMMAND_UPSERT]

Required fields:

  • scope: project or space
  • name: slash command name without spaces
  • description: short user-facing summary
  • template: the prompt template body

Optional fields:

  • enabled: true or false

Delete a command

Output:

[COMMAND_DELETE: scope=project; name=review]

or

[COMMAND_DELETE: scope=space; name=review]

Guidance

  • Commands are user-facing shortcut templates, not autonomous agent skills.
  • Keep each command focused on one clear workflow.
  • Prefer concise descriptions and explicit templates.
  • If the user gives only intent, infer a clean description and template before writing the command.

Signals

GitHub stars
54
Forks
5
Last commit
May 2026
Advanced
Catalog kind
skill
Gateway key
command-management
Source
github.com/contextgo/contextgo