Skip to main content

/can - Capability Check

Check if Claude, an agent, skill, tool, or system has a specific capability.

System Prompt

⚠️ EXECUTION DIRECTIVE: When the user invokes this command, you MUST:

  1. IMMEDIATELY execute - no questions, no explanations first
  2. Parse the query to identify subject and capability
  3. Check capability against known features/tools/skills
  4. Answer definitively with YES/NO plus explanation

DO NOT:

  • Give vague "it depends" answers without specifics
  • Skip execution even if query seems obvious

The user invoking the command IS the confirmation.


Usage

/can <subject> <capability>

Query Patterns

Query PatternPurposeLookup Method
/can claude <action>Check Claude capabilitiesBuilt-in knowledge + tool list
/can agent <name> <action>Check agent capabilitiesAgent frontmatter tools: field
/can skill <name> <action>Check skill capabilitiesSkill SKILL.md documentation
/can this codebase <action>Check project capabilitiesPackage.json, requirements.txt
/can <tool> <action>Check tool capabilitiesTool documentation lookup

Execution Steps

Step 1: Parse Query

# Extract subject and capability
SUBJECT=$(echo "$QUERY" | awk '{print $1}')
CAPABILITY=$(echo "$QUERY" | cut -d' ' -f2-)

Step 2: Capability Lookup

# For Claude capabilities
if [ "$SUBJECT" = "claude" ]; then
# Check against known tool list and features
# Read, Write, Edit, Bash, Grep, Glob, WebFetch, WebSearch, Task, etc.
fi

# For agent capabilities
if [ "$SUBJECT" = "agent" ]; then
AGENT_NAME=$(echo "$CAPABILITY" | awk '{print $1}')
# Read agent file and check tools: field
grep -A 10 "^tools:" "agents/$AGENT_NAME.md"
fi

# For skill capabilities
if [ "$SUBJECT" = "skill" ]; then
SKILL_NAME=$(echo "$CAPABILITY" | awk '{print $1}')
# Read skill and check documented features
cat "skills/$SKILL_NAME/SKILL.md"
fi

Step 3: Respond with Answer

# Format: YES/NO + explanation + alternatives if NO

Response Format

┌─────────────────────────────────────────────────────────────────┐
│ /can: <query> │
├─────────────────────────────────────────────────────────────────┤
│ CAPABILITY CHECK │
│ ──────────────── │
│ │
│ ✅ YES / ❌ NO │
│ │
│ Subject: <subject> │
│ Capability: <capability> │
│ │
│ Explanation: │
│ <detailed explanation of capability or limitation> │
│ │
│ [If NO] Alternatives: │
│ • <alternative approach 1> │
│ • <alternative approach 2> │
│ │
├─────────────────────────────────────────────────────────────────┤
│ Source: <tool list | agent frontmatter | skill docs> │
└─────────────────────────────────────────────────────────────────┘

Examples

Check Claude Capability

/can claude read PDFs

Output:

✅ YES
Claude can read PDF files using the Read tool. For PDFs over 10 pages,
use the `pages` parameter to specify page ranges (e.g., pages: "1-5").
Maximum 20 pages per request.

Check Agent Capability

/can agent devops-engineer deploy to kubernetes

Output:

✅ YES
The devops-engineer agent has Bash tool access and kubernetes-patterns
skill. It can execute kubectl commands and manage K8s deployments.
Tools: Read, Write, Edit, Bash, Grep, Glob

Check Skill Capability

/can skill pptx create animations

Output:

❌ NO
The pptx skill converts HTML to PPTX via LibreOffice. Animations are
not preserved in this conversion workflow.

Alternatives:
• Use Google Slides API for animation support
• Create animations in PowerPoint post-generation

Check Codebase Capability

/can this codebase handle GraphQL

Output:

❌ NO (currently)
No GraphQL dependencies found in package.json or requirements.txt.

To add GraphQL:
• Python: `pip install graphene-django`
• JavaScript: `npm install @apollo/server`
CommandPurpose
/who <component>Find who owns/maintains something
/what <component>Discover what a component is
/which <task>Find which agent/skill to use
/does <component> <action>Verify if something does an action
/is <state>Check current state

Success Output

✅ COMMAND COMPLETE: /can
Query: <query>
Answer: YES/NO
Subject: <subject>
Capability: <capability>

Completion Checklist

Before marking complete:

  • Subject identified
  • Capability parsed
  • Definitive YES/NO given
  • Explanation provided
  • Alternatives offered (if NO)

Failure Indicators

This command has FAILED if:

  • ❌ No query provided
  • ❌ Subject not identifiable
  • ❌ Vague "maybe" answer given
  • ❌ No explanation provided

When NOT to Use

Do NOT use when:

  • Need step-by-step instructions (use /how)
  • Want recommendations (use /which)
  • Need state verification (use /is)

Principles

This command embodies:

  • #6 Clear, Understandable - Definitive YES/NO answers
  • #9 Based on Facts - Uses actual tool/feature documentation
  • #3 Complete Execution - Always provides explanation

Full Standard: CODITECT-STANDARD-AUTOMATION.md


Version: 1.0.0 Created: 2026-02-04 Author: CODITECT Team Framework: 5W+H Question Framework