claude-code - 💡(How to fix) Fix [DOCS] Commands reference missing `/less-permission-prompts` bundled skill [1 participants]

Documentation Type

Missing documentation (feature not documented)

Documentation Location

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

Section/Topic

Bundled skills in the commands reference, plus the related permissions/settings workflow for generating allow rules from transcript history

Current Documentation

The commands reference currently says:

"The table below lists all the commands included in Claude Code. Entries marked Skill are bundled skills. They use the same mechanism as skills you write yourself: a prompt handed to Claude, which Claude can also invoke automatically when relevant."

The same table lists bundled skills such as /batch, /claude-api, /debug, /loop, and /simplify, but there is no /less-permission-prompts entry.

The skills page currently says:

"Claude Code includes a set of bundled skills that are available in every session, including /simplify, /batch, /debug, /loop, and /claude-api."

The permissions docs explain how allow rules work:

"You can view and manage Claude Code's tool permissions with /permissions. This UI lists all permission rules and the settings.json file they are sourced from."

The settings docs show settings.json permission allowlists, for example:

"permissions": { "allow": [ "Bash(npm run lint)", "Bash(npm run test *)", "Read(~/.zshrc)" ] }

The Claude directory docs also say:

"Anything that passes through a tool lands in a transcript on disk: file contents, command output, pasted text."

No documentation currently exists for /less-permission-prompts, what it scans, or how its proposed allowlist should be reviewed and applied.

What's Wrong or Missing?

Changelog v2.1.111 added /less-permission-prompts, described as a skill that scans transcripts for common read-only Bash and MCP tool calls and proposes a prioritized allowlist for .claude/settings.json, but the docs do not currently cover this feature.

Missing pieces include:

A. Discovery

Users cannot discover that /less-permission-prompts exists from the commands reference or bundled skills docs.

B. Behavior

There is no explanation of what transcript data the skill scans, what counts as a common read-only Bash or MCP call, or what output the skill returns.

C. Configuration workflow

The docs explain permission allow rules in general, but not how this skill helps generate a prioritized allowlist for .claude/settings.json.

D. Safety/privacy guidance

Because transcript files can contain tool inputs and outputs, the docs should explain any review expectations and privacy considerations before using transcript-derived allowlist suggestions.

Suggested Improvement

Add /less-permission-prompts to the commands reference and bundled skills documentation, then add a short task-oriented subsection that explains:

1. How to invoke the skill
2. That it analyzes local transcript history for repeated read-only Bash and MCP permission prompts
3. What kind of prioritized recommendations it returns for permissions.allow
4. How to review and copy those recommendations into .claude/settings.json
5. When to use ~/.claude/settings.json versus project .claude/settings.json
6. Any privacy caveats related to scanning local transcripts

If the skill emits a JSON snippet or suggested rule list, include one concrete example so users can see the expected output format.

Impact

High - Prevents users from using a feature

Additional Context

Affected Pages:

Total scope: 5 pages affected

Source: Changelog v2.1.111

Exact changelog entry text:

Added /less-permission-prompts skill — scans transcripts for common read-only Bash and MCP tool calls and proposes a prioritized allowlist for .claude/settings.json