laststance/skill-inspect

Skill Inspector

Codex Compatibility

When running this skill in Codex, translate Claude Code-only primitives before acting: AskUserQuestion -> chat/request_user_input, TodoWrite -> update_plan, Task/TaskCreate/TeamCreate/SendMessage -> spawn_agent/send_input/wait_agent when available and allowed, and EnterPlanMode/ExitPlanMode -> a concise chat plan plus explicit approval. Resolve Read/Write/Edit/Bash/WebSearch/WebFetch to Codex file/shell/web tools, and map ~/.claude/... paths to ~/.agents/... or ~/.codex/... unless the task explicitly targets Claude Code.

Cursor Compatibility

When running this skill in Cursor Agent, translate Claude Code-only primitives before acting: AskUserQuestion -> AskQuestion; TodoWrite -> Cursor TodoWrite or an equivalent checklist; Task/TaskCreate/TeamCreate/SendMessage/multi-agent flows -> Cursor Task (subagents), parallel Tasks, or run_in_background when allowed (TeamCreate/SendMessage may have no exact match); EnterPlanMode/ExitPlanMode -> Plan mode (SwitchMode / CreatePlan) plus explicit user approval. Resolve Read/Write/Edit/StrReplace/Bash/web/search/MCP via Cursor Composer or Agent equivalents. MCP names written as mcp__server__tool typically map to call_mcp_tool with configured server identifiers. Map ~/.claude/... to ~/.cursor/skills/, .cursor/skills/, and .cursor/rules/ unless the task explicitly targets Claude Code.

Read-only diagnostic. Resolves any name across the skill ecosystem and displays a structured info card with provenance, metadata, and cross-tool availability.

<essential_principles>

• Read-only: never modify, install, or execute anything
• Report ALL matches across all categories (a name may exist in multiple places)
• Always include cross-tool availability check for skills

</essential_principles>

Step 1: Resolve Name

Extract <name> from the user's argument. Run all checks — report every hit:

1a. Skill (~/.claude/skills/ + ~/.agents/skills/)

test -d ~/.claude/skills/<name> && echo "SKILL_FOUND_CLAUDE"
test -d ~/.agents/skills/<name> && echo "SKILL_FOUND_AGENTS"

If found in either location, sub-classify:

if [ -L ~/.claude/skills/<name> ]; then
  target=$(readlink ~/.claude/skills/<name>)
  echo "SYMLINK: $target"
elif [ -d ~/.claude/skills/<name> ]; then
  echo "REAL_DIR"
fi

Then Read ~/.agents/.skill-lock.json and check if <name> is a key in .skills:

| Condition | Type Label | |-----------|------------| | Symlink AND in skill-lock.json | CLI-installed (npx skills) | | Symlink target contains gstack/ | gstack (bundled) | | Real directory, NOT in skill-lock.json | Hand-crafted (local) | | Only in ~/.agents/skills/ (no symlink in ~/.claude/skills/) | CLI-installed (not linked to Claude Code) |

1b. Plugin Skill

Read ~/.claude/plugins/installed_plugins.json to get all plugin installPath values.

For each plugin, check:

test -d {installPath}/skills/<name> && echo "PLUGIN_SKILL_FOUND"

If found, extract from installed_plugins.json: plugin name (key), version, installedAt, marketplace (from key format plugin@marketplace).

1c. Legacy Command

test -f ~/.claude/commands/<name>.md && echo "LEGACY_COMMAND"

1d. Agent Definition

test -f ~/.claude/agents/<name>.md && echo "AGENT_FOUND"

1e. MCP Server

Read ~/.claude.json and check if <name> is a key in mcpServers. Also Read ~/.claude/.mcp.json and check mcpServers.

If found, note: "MCP Server is a tool provider, not a skill."

1f. Built-in Command

Check against the list in @references/built-ins.md.

1g. No Match — Fuzzy Search

If nothing found:

ls ~/.claude/skills/ ~/.agents/skills/ ~/.claude/commands/ ~/.claude/agents/ 2>/dev/null | grep -i "<name>"

Suggest up to 5 closest substring matches.

Step 2: Gather Metadata

Based on resolved type, collect fields:

| Type | Read | Fields | |------|------|--------| | CLI-installed | SKILL.md frontmatter + skill-lock.json entry | description, argument-hint, source, installedAt, updatedAt | | Hand-crafted | SKILL.md frontmatter + git log --format='%ai' -1 -- ~/.claude/skills/<name> | description, argument-hint, last git commit date | | gstack | SKILL.md frontmatter + readlink target | description, argument-hint, bundle path | | Plugin Skill | SKILL.md frontmatter + installed_plugins.json | description, plugin@marketplace, version, installedAt | | Legacy Command | .md frontmatter | description | | Agent Definition | .md frontmatter | description, model | | MCP Server | config JSON | transport type (stdio/http), command or url | | Built-in | @references/built-ins.md | description |

Step 3: Cross-Tool Availability

For any skill found, check presence across AI Agent tools:

for dir in ~/.agents/skills ~/.claude/skills ~/.cursor/skills ~/.codex/skills ~/.gemini/skills ~/.vscode/skills ~/.antigravity/skills; do
  tool=$(basename "$(dirname "$dir")")
  tool=${tool#.}
  [ -d "$dir/<name>" ] && echo "✅ $tool" || echo "— $tool"
done

Step 4: Display Info Card

Present using this format:

## 🔍 <name>

| Field | Value |
|-------|-------|
| **Type** | <type label> |
| **Description** | <from frontmatter> |
| **Location** | <full path> |
| **Source** | <repo / plugin@marketplace / "local"> |
| **Installed** | <YYYY-MM-DD or "—"> |
| **Updated** | <YYYY-MM-DD or "—"> |
| **Arguments** | <argument-hint or "—"> |

### Available In
✅ claude  ✅ cursor  ✅ codex  ✅ gemini  — vscode  — antigravity

### Capabilities
<First content section of SKILL.md body after frontmatter, up to ~10 lines>

Adapt fields per type:

• CLI-installed / Hand-crafted / gstack / Plugin: show all fields
• Legacy Command: omit Source, Installed, Updated; show "⚠️ Deprecated — migrate to Skills format"
• Agent Definition: show model field instead of Arguments; omit Available In
• MCP Server: show transport + connection info; omit Available In; add note
• Built-in: show description only; note "Hardcoded in Claude Code application"

If multiple matches found, display one card per match with a separator.

Boundaries

Will: Read files, check symlinks, display info, suggest fuzzy matches. Will Not: Modify files, install/uninstall, execute skills, query running MCP servers.

@references/built-ins.md