Agent skill
permissions
Interactive Claude Code permissions management - scans plugin recommendations, presents categorized setup, and writes to settings.json
Install this agent skill to your Project
npx add-skill https://github.com/oakensoul/aida-core-plugin/tree/main/skills/permissions
SKILL.md
Permissions Skill
Manages Claude Code permissions interactively by scanning installed
plugin recommendations, presenting categorized choices, and writing
the selected configuration to settings.json.
Activation
This skill activates when /aida config permissions is invoked
through the aida skill routing system.
Command Routing
| Command | Description |
|---|---|
/aida config permissions |
Interactive permissions setup |
/aida config permissions --audit |
Audit current permissions |
Path Resolution
Scripts are located at skills/permissions/scripts/ relative to
the plugin root.
- Plugin cache:
~/.claude/plugins/cache/*/*/.claude-plugin/ - User settings:
~/.claude/settings.json - Project settings:
.claude/settings.json - Local settings:
.claude/settings.local.json
Two-Phase API
Phase 1: get_questions(context)
- Scans installed plugins for
recommendedPermissionsinaida-config.json - Deduplicates and categorizes discovered rules
- Reads current permissions from all settings scopes
- Detects conflicts between current and proposed rules
- Returns questions for preset selection and per-category allow/ask/deny choices
Phase 2: execute(context, responses)
- Applies the selected preset or custom category choices
- Builds the final rules dictionary (allow/ask/deny lists)
- Writes permissions to the chosen settings scope
- Returns a summary of changes made
Audit Mode
When --audit is passed, the skill skips the interactive flow
and returns a coverage/gap/conflict analysis report.
Resources
Scripts
| Script | Purpose |
|---|---|
scripts/permissions.py |
Two-phase API entry point |
scripts/scanner.py |
Plugin permission discovery |
scripts/aggregator.py |
Deduplication and categorization |
scripts/settings_manager.py |
Settings.json read/write |
References
| Document | Purpose |
|---|---|
references/permissions-workflow.md |
End-to-end workflow guide |
references/rule-syntax.md |
Permission rule format docs |
references/presets.md |
Preset definitions and mappings |
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
agent-manager
Manages Claude Code agent (subagent) definitions including create, validate, version, and list operations using a two-phase API.
plugin-manager
Unified plugin management combining extension CRUD operations (create, validate, version, list) with project scaffolding
skill-manager
Manages Claude Code skill definitions - create, validate, version, and list skills following the Agent Skills open standard.
hook-manager
Manages Claude Code hook configurations in settings.json files. Supports listing, adding, removing, and validating lifecycle hooks.
aida
This skill routes /aida commands to appropriate handlers - configuration, diagnostics, feedback, extension management (agent-manager, skill-manager, plugin-manager, hook-manager, claude-md-manager), and session persistence (memento).
project-context
Provides current project state, architecture overview, and structured metadata
Didn't find tool you were looking for?