CowboyLogic/ai-dev / skills/client-config-claudecode
client-config-claudecode
FlaggedManage and maintain Claude Code configuration files and settings. Use this skill whenever the user wants to view, edit, add, remove, or understand any Claude Code settings — permissions, hooks, MCP servers, environment variables, model settings, sandbox, auto mode, or any other configuration field. Trigger on: "add permission", "allow tool", "deny tool", "configure hook", "add MCP server", "change model", "update settings", "show my settings", "what settings do I have", or "set default mode".
| File | Kind | Size |
|---|---|---|
| CHANGELOG.md | markdown | 18.3 KB |
| README.md | markdown | 3.2 KB |
| SKILL.md | skill md | 8.5 KB |
| assets/sources.json | config | 3.2 KB |
| evals/evals.json | config | 10.2 KB |
| references/hooks.md | markdown | 28.0 KB |
| references/mcp.md | markdown | 16.1 KB |
| references/permissions.md | markdown | 21.6 KB |
| references/settings-schema.md | markdown | 39.2 KB |
| scripts/show-settings.py | script | 8.6 KB |
| scripts/update-references.py | script | 5.4 KB |
| scripts/validate-settings.py | script | 35.9 KB |
evals/evals.json
View on GitHub{
"skill_name": "client-config-claudecode",
"evals": [
{
"id": 1,
"name": "hook-lint-on-edit",
"prompt": "Add a hook to my Claude Code settings so that whenever Claude edits or writes any file, it automatically runs `npm run lint --fix 2>&1` in the project root. Show me the exact JSON to add to my ~/.claude/settings.json.",
"expected_output": "A complete, correct hooks configuration showing PostToolUse event with Edit|Write matcher and a command handler running npm run lint --fix",
"assertions": [
{
"id": "uses-PostToolUse-event",
"description": "Response uses the PostToolUse hook event (not PreToolUse)"
},
{
"id": "uses-edit-write-matcher",
"description": "Matcher covers both Edit and Write tools"
},
{
"id": "includes-project-dir-var",
"description": "Uses $CLAUDE_PROJECT_DIR or equivalent to run in project root"
},
{
"id": "valid-json-structure",
"description": "The JSON snippet is syntactically valid and uses the correct hooks nesting structure"
},
{
"id": "no-hallucinated-fields",
"description": "Does not invent non-existent hook fields or events"
}
]
},
{
"id": 2,
"name": "permissions-git-allow-deny",
"prompt": "Configure my Claude Code settings to allow git read commands (status, log, diff, show) but block destructive git operations like push, reset --hard, and force-push. Show me the exact permissions config.",
"expected_output": "Correct allow rules for safe git read commands and deny rules for destructive git operations using proper Bash() rule syntax",
"assertions": [
{
"id": "allow-git-read-commands",
"description": "Includes allow rules covering git status, git log, git diff, and/or git show"
},
{
"id": "deny-git-push",
"description": "Includes a deny rule that blocks git push"
},
{
"id": "deny-git-reset-hard",
"description": "Includes a deny rule that blocks git reset --hard or git reset"
},
{
"id": "correct-bash-rule-syntax",
"description": "Uses Bash(...) rule syntax correctly (not just strings like 'git push')"
},
{
"id": "deny-before-allow-awareness",
"description": "Response shows awareness that deny takes precedence over allow (evaluation order)"
}
]
},
{
"id": 3,
"name": "voice-and-model-config",
"prompt": "I want to enable voice dictation in tap mode with auto-submit turned on, and set my default model to claude-sonnet-4-6. Show me the complete config snippet for my settings.json.",
"expected_output": "Correct voice object (not deprecated voiceEnabled) plus model setting in valid JSON, noting that voice.autoSubmit only applies in hold mode",
"assertions": [
{
"id": "uses-voice-object-not-deprecated",
"description": "Uses the voice object {enabled, mode, autoSubmit} rather than the deprecated voiceEnabled boolean"
},
{
"id": "voice-mode-tap",
"description": "Sets voice.mode to tap"
},
{
"id": "voice-autosubmit-true",
"description": "Sets voice.autoSubmit to true"
},
{
"id": "autosubmit-hold-only-caveat",
"description": "Notes that voice.autoSubmit applies in hold mode only, so it has no effect with mode tap"
},
{
"id": "model-key-correct",
"description": "Uses the model key (not modelId or defaultModel) set to claude-sonnet-4-6"
},
{
"id": "valid-json",
"description": "The snippet is valid JSON that could be merged into settings.json"
}
]
},
{
"id": 4,
"name": "default-mode-scope",
"prompt": "I want every session in this repo to start in bypassPermissions mode. Add it to the project's .claude/settings.json for me.",
"expected_output": "Explains that defaultMode auto/bypassPermissions don't take effect from project or local settings (v2.1.257+) and that it must go in ~/.claude/settings.json, with a safety caution about bypassPermissions",
"assertions": [
{
"id": "project-scope-ignored",
"description": "States that bypassPermissions as defaultMode is ignored in project/local settings"
},
{
"id": "user-settings-location",
"description": "Directs the setting to ~/.claude/settings.json (or --permission-mode / --dangerously-skip-permissions for one session)"
},
{
"id": "safety-caution",
"description": "Warns that bypassPermissions should only be used in isolated environments"
},
{
"id": "valid-json",
"description": "Any JSON shown uses permissions.defaultMode with valid syntax"
}
]
},
{
"id": 5,
"name": "hook-fail-closed",
"prompt": "My PreToolUse hook in .claude/settings.json runs a policy script before every Bash command. I noticed that when the script path was mistyped, Bash commands still ran. Make the hook block the command when the script fails instead.",
"expected_output": "Sets onFailure: \"block\" on the command hook (v2.1.295+), explains that a hook that can't start, times out, or exits with a code other than 0 or 2 otherwise lets the action through, and notes exit 2 is the blocking exit code",
"assertions": [
{
"id": "on-failure-block",
"description": "Adds \"onFailure\": \"block\" to the command hook handler"
},
{
"id": "default-non-blocking",
"description": "Explains that without it a failed, missing, or timed-out hook is a non-blocking error and the action proceeds"
},
{
"id": "version-gate",
"description": "Notes that onFailure requires Claude Code v2.1.295 or later"
},
{
"id": "exit-2-blocks",
"description": "Notes that exit code 2 is the blocking exit code, so a script that exits 2 blocks the action without onFailure"
},
{
"id": "valid-hook-shape",
"description": "Hook JSON nests the handler inside a matcher group under hooks.PreToolUse with a valid type"
}
]
},
{
"id": 6,
"name": "stop-hook-loop-cap",
"prompt": "I have a Stop hook that blocks Claude from finishing until the test suite passes. It works, but after a while Claude just stops even though the tests are still failing. Why does that happen, and how do I let it keep trying longer?",
"expected_output": "Explains Claude Code's cap of 8 consecutive stop-hook continuations, that the count resets when Claude calls a tool, and that CLAUDE_CODE_STOP_HOOK_BLOCK_CAP raises it (0 disables it), set in the env block of settings.json",
"assertions": [
{
"id": "eight-cap",
"description": "States that after 8 consecutive continuations from stop hooks Claude Code overrides the next block and ends the turn"
},
{
"id": "counter-resets",
"description": "Notes that the count resets each time Claude calls a tool"
},
{
"id": "block-cap-variable",
"description": "Names CLAUDE_CODE_STOP_HOOK_BLOCK_CAP as the way to raise the cap, with 0 disabling it"
},
{
"id": "stop-hook-active",
"description": "Recommends the hook check stop_hook_active so it does not block on a condition that never resolves"
},
{
"id": "env-block",
"description": "If it shows configuration, sets the variable under the env key of settings.json with valid JSON"
}
]
},
{
"id": 7,
"name": "idle-compaction",
"prompt": "Claude Code keeps compacting my long session when I step away for a while and come back. I want to turn that off but keep normal auto-compact working. What do I set?",
"expected_output": "Sets idleCompaction to false in settings.json (any settings file, v2.1.290+), explains it compacts an idle 200K+ token conversation shortly before the one-hour prompt cache expires, and notes autoCompactEnabled is the separate key for normal auto-compact",
"assertions": [
{
"id": "idle-key",
"description": "Names the idleCompaction setting and sets it to false"
},
{
"id": "keeps-auto-compact",
"description": "Leaves autoCompactEnabled alone, or says it controls normal auto-compact and that turning it off also turns idle compaction off"
},
{
"id": "min-version",
"description": "Notes idleCompaction needs Claude Code v2.1.290 or later"
},
{
"id": "threshold",
"description": "Mentions the alternative of raising CLAUDE_CODE_IDLE_COMPACT_MIN_TOKENS (default 200000, floor 100000) instead of turning it off"
}
]
},
{
"id": 8,
"name": "mcp-server-name-with-double-underscore",
"prompt": "I have an MCP server configured as reports__v2. I added \"mcp__reports__v2\" to permissions.allow in settings.json but Claude still prompts for every one of its tools. What's wrong?",
"expected_output": "Explains that the bare mcp__reports__v2 rule is not read as a rule for the reports__v2 server: under the mcp__<server>__<tool> grammar it names server reports (and tool v2), so it does not match the reports__v2 server's tools, and gives mcp__reports__v2__* as the rule to write",
"assertions": [
{
"id": "misread-rule",
"description": "Explains that mcp__reports__v2 (and mcp__reports) is parsed against a server named reports (for mcp__reports__v2, with v2 as a tool name) and so does not allow the reports__v2 server's tools"
},
{
"id": "wildcard-form",
"description": "Gives mcp__reports__v2__* (the wildcard form with the whole server name) as the rule to write"
},
{
"id": "name-ends-underscore",
"description": "Notes the same wildcard form is needed for a server name that ends in an underscore"
}
]
}
]
}
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |