claude-code - 💡(How to fix) Fix [DOCS] Skill tool docs still say built-in slash commands cannot be invoked [1 participants]

Documentation Type

Incorrect/outdated documentation

Documentation Location

https://code.claude.com/docs/en/skills

Section/Topic

Restrict Claude's skill access and the duplicated built-in-command guidance on slash-commands

Current Documentation

The docs currently say on both skills and slash-commands:

By default, Claude can invoke any skill that doesn't have disable-model-invocation: true set. Skills that define allowed-tools grant Claude access to those tools without per-use approval when the skill is active. Your permission settings still govern baseline approval behavior for all other tools. Built-in commands like /compact and /init are not available through the Skill tool.

The command reference also shows the built-in commands affected by this change:

/init | Initialize project with a CLAUDE.md guide.

/review | Deprecated. Install the code-review plugin instead.

/security-review | Analyze pending changes on the current branch for security vulnerabilities.

What's Wrong or Missing?

Changelog v2.1.108 says:

The model can now discover and invoke built-in slash commands like /init, /review, and /security-review via the Skill tool.

That directly contradicts the current Skill tool documentation, which still says built-in commands like /init are not available through the Skill tool.

As written, the docs give users the wrong mental model for how built-in slash commands interact with the Skill tool:

• they imply built-in slash commands are excluded from model discovery and invocation
• they do not explain that built-in commands such as /init, /review, and /security-review are now part of the Skill-tool-discoverable command surface
• they make the permission and invocability guidance in skills / slash-commands outdated after v2.1.108

Suggested Improvement

Update the Skill tool documentation in both skills and slash-commands to reflect the v2.1.108 behavior.

Minimum fix:

1. Replace the blanket statement that built-in commands like /compact and /init are unavailable through the Skill tool.
2. Explicitly document that the model can now discover and invoke built-in slash commands such as /init, /review, and /security-review via the Skill tool.
3. If some built-in commands are still excluded, document the real exceptions explicitly instead of grouping /init with them.
4. Add a short note near the permission guidance explaining that this change affects how users should reason about Skill-tool access to built-in slash commands.

Impact

Medium - Makes feature difficult to understand

Additional Context

Affected Pages:

Total scope: 3 pages affected

Source: Changelog v2.1.108

Exact changelog entry:

The model can now discover and invoke built-in slash commands like /init, /review, and /security-review via the Skill tool