As a developer, I want shell scripts used by the skill to run without repeated permission prompts so that skill execution is smooth

[kiyotis]

Situation

During skill execution, scripts such as full-text-search.sh and read-sections.sh are called multiple times. Each execution triggers a Claude Code permission prompt asking the user to allow or deny the Bash command.

A single code analysis run may invoke these scripts 5-10 times, resulting in the same number of interruptions.

Pain

Users must repeatedly approve the same shell script executions during a single skill run, which is disruptive and adds friction to what should be an automated workflow.

Benefit

• Users can run the skill without manual interruptions for script permission prompts
• Skill execution feels automated and smooth

Success Criteria

• Shell scripts used by nabledge-6 skill are added to the allow list in .claude/settings.json via the setup script
• Setup script (setup-6-cc.sh) writes the necessary permission entries
• Users who already installed the skill can re-run setup to apply the permission entries
• Permitted script paths are documented