Investigate logo

Investigate

Organization
MadAppGang
investigate

Unified entry point for code investigation. Auto-routes to specialized detective based on query keywords. Use when investigation type is unclear or for general exploration.

Overview

PublisherMadAppGang
Repositoryclaude-code
Skill nameinvestigate
Stars
281
Forks
26
Bundled files
Instructions only
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by MadAppGang on GitHub. Read the source before you install it.

Installation

Install the Investigate AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/MadAppGang/claude-code.git /tmp/claude-code
mkdir -p .claude/skills
cp -r /tmp/claude-code/plugins/code-analysis/skills/investigate .claude/skills/investigate
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Investigate in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Investigate on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Investigate is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Investigate Skill

Version: 1.0.0 Purpose: Keyword-based routing to specialized detective skills Pattern: Smart delegation via Task tool

Overview

This skill analyzes your investigation query and routes to the appropriate detective specialist:

  • debugger-detective (errors, bugs, crashes)
  • tester-detective (tests, coverage, edge cases)
  • architect-detective (architecture, design, patterns)
  • developer-detective (implementation, data flow - default)

Routing Logic

Priority System (Highest First)

  1. Error/Debug (Priority 1) - Time-critical bug fixes

    • Keywords: "debug", "error", "broken", "failing", "crash"
    • Route to: debugger-detective
  2. Testing (Priority 2) - Specialized test analysis

    • Keywords: "test", "coverage", "edge case", "mock"
    • Route to: tester-detective
  3. Architecture (Priority 3) - High-level understanding

    • Keywords: "architecture", "design", "structure", "layer"
    • Route to: architect-detective
  4. Implementation (Default, Priority 4) - Most common

    • Keywords: "implementation", "how does", "code flow"
    • Route to: developer-detective

Conflict Resolution

When multiple keywords from different categories are detected:

  • Highest priority wins (Priority 1 beats Priority 2, etc.)
  • No matches: Default to developer-detective

Workflow

Phase 1: Extract Query

The investigation query should be available from the task description or user input.

bash
# Query comes from the Task description or user request
INVESTIGATION_QUERY="${TASK_DESCRIPTION:-$USER_QUERY}"

# Normalize to lowercase for case-insensitive matching
QUERY_LOWER=$(echo "$INVESTIGATION_QUERY" | tr '[:upper:]' '[:lower:]')

Phase 2: Keyword Detection

bash
# Priority 1: Error/Debug keywords
if echo "$QUERY_LOWER" | grep -qE "debug|error|broken|failing|crash"; then
  DETECTIVE="debugger-detective"
  KEYWORDS="debug/error keywords"
  PRIORITY=1
  RATIONALE="Bug fixes are time-critical and require call chain tracing"

# Priority 2: Testing keywords
elif echo "$QUERY_LOWER" | grep -qE "test|coverage|edge case|mock"; then
  DETECTIVE="tester-detective"
  KEYWORDS="test/coverage keywords"
  PRIORITY=2
  RATIONALE="Test analysis is specialized and requires callers analysis"

# Priority 3: Architecture keywords
elif echo "$QUERY_LOWER" | grep -qE "architecture|design|structure|layer"; then
  DETECTIVE="architect-detective"
  KEYWORDS="architecture/design keywords"
  PRIORITY=3
  RATIONALE="High-level understanding requires PageRank analysis"

# Priority 4: Implementation (default)
else
  DETECTIVE="developer-detective"
  KEYWORDS="implementation (default)"
  PRIORITY=4
  RATIONALE="Most common investigation type - data flow via callers/callees"
fi

Phase 3: User Feedback

Before delegating, inform the user of the routing decision:

bash
echo ""
echo "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━"
echo "🔍 Investigation Routing"
echo "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━"
echo ""
echo "Query: $INVESTIGATION_QUERY"
echo ""
echo "Detected: $KEYWORDS (Priority $PRIORITY)"
echo "Routing to: $DETECTIVE"
echo "Reason: $RATIONALE"
echo ""
echo "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━"
echo ""

Phase 4: Delegation via Task Tool

Use the Task tool to delegate to the selected detective:

typescript
Task({
  description: INVESTIGATION_QUERY,
  agent: DETECTIVE,
  context: {
    routing_reason: `Auto-routed based on ${KEYWORDS}`,
    original_query: INVESTIGATION_QUERY,
    priority: PRIORITY
  }
})

Examples

Example 1: Debug Keywords

Input: "Why is login broken?"

Detection:

  • Keyword matched: "broken"
  • Priority: 1 (Error/Debug)
  • Route to: debugger-detective

Feedback:

🔍 Investigation Routing
Query: Why is login broken?
Detected: debug/error keywords (Priority 1)
Routing to: debugger-detective
Reason: Bug fixes are time-critical and require call chain tracing

Example 2: Test Keywords

Input: "What's the test coverage for payment?"

Detection:

  • Keywords matched: "test", "coverage"
  • Priority: 2 (Testing)
  • Route to: tester-detective

Feedback:

🔍 Investigation Routing
Query: What's the test coverage for payment?
Detected: test/coverage keywords (Priority 2)
Routing to: tester-detective
Reason: Test analysis is specialized and requires callers analysis

Example 3: Architecture Keywords

Input: "What's the architecture of the auth layer?"

Detection:

  • Keywords matched: "architecture", "layer"
  • Priority: 3 (Architecture)
  • Route to: architect-detective

Feedback:

🔍 Investigation Routing
Query: What's the architecture of the auth layer?
Detected: architecture/design keywords (Priority 3)
Routing to: architect-detective
Reason: High-level understanding requires PageRank analysis

Example 4: No Keywords (Default)

Input: "How does payment work?"

Detection:

  • No keywords matched
  • Priority: 4 (Default)
  • Route to: developer-detective

Feedback:

🔍 Investigation Routing
Query: How does payment work?
Detected: implementation (default) (Priority 4)
Routing to: developer-detective
Reason: Most common investigation type - data flow via callers/callees

Example 5: Multi-Keyword Conflict

Input: "Debug the test coverage"

Detection:

  • Keywords matched: "debug" (Priority 1) AND "test" (Priority 2)
  • Priority 1 wins
  • Route to: debugger-detective

Feedback:

🔍 Investigation Routing
Query: Debug the test coverage
Detected: debug/error keywords (Priority 1)
Routing to: debugger-detective
Reason: Bug fixes are time-critical and require call chain tracing
(Note: Also detected test keywords, but debug takes priority)

Complete Implementation

Here's the full workflow:

bash
#!/bin/bash

# Get investigation query from task description
INVESTIGATION_QUERY="${TASK_DESCRIPTION}"

# Normalize to lowercase
QUERY_LOWER=$(echo "$INVESTIGATION_QUERY" | tr '[:upper:]' '[:lower:]')

# Keyword detection with priority routing
if echo "$QUERY_LOWER" | grep -qE "debug|error|broken|failing|crash"; then
  DETECTIVE="debugger-detective"
  KEYWORDS="debug/error keywords"
  PRIORITY=1
  RATIONALE="Bug fixes are time-critical and require call chain tracing"

elif echo "$QUERY_LOWER" | grep -qE "test|coverage|edge case|mock"; then
  DETECTIVE="tester-detective"
  KEYWORDS="test/coverage keywords"
  PRIORITY=2
  RATIONALE="Test analysis is specialized and requires callers analysis"

elif echo "$QUERY_LOWER" | grep -qE "architecture|design|structure|layer"; then
  DETECTIVE="architect-detective"
  KEYWORDS="architecture/design keywords"
  PRIORITY=3
  RATIONALE="High-level understanding requires PageRank analysis"

else
  DETECTIVE="developer-detective"
  KEYWORDS="implementation (default)"
  PRIORITY=4
  RATIONALE="Most common investigation type - data flow via callers/callees"
fi

# Show routing decision
echo ""
echo "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━"
echo "🔍 Investigation Routing"
echo "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━"
echo ""
echo "Query: $INVESTIGATION_QUERY"
echo ""
echo "Detected: $KEYWORDS (Priority $PRIORITY)"
echo "Routing to: $DETECTIVE"
echo "Reason: $RATIONALE"
echo ""
echo "━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━"
echo ""

Then use the Task tool to delegate:

typescript
Task({
  description: INVESTIGATION_QUERY,
  agent: DETECTIVE
})

Fallback Protocol

If routing produces unexpected results:

  1. Show routing decision to user
  2. Ask for override if needed via AskUserQuestion
  3. Default to developer-detective if ambiguous

Override Pattern

typescript
// If user wants to override the routing
AskUserQuestion({
  questions: [{
    question: `Auto-routing selected ${DETECTIVE}. Override?`,
    header: "Investigation Routing",
    multiSelect: false,
    options: [
      { label: "Continue with auto-routing", description: `Use ${DETECTIVE}` },
      { label: "debugger-detective", description: "Root cause analysis" },
      { label: "tester-detective", description: "Test coverage analysis" },
      { label: "architect-detective", description: "Architecture patterns" },
      { label: "developer-detective", description: "Implementation details" }
    ]
  }]
})

Integration with Existing Workflow

This skill is additive only and does not change existing behavior:

  • Direct detective usage still works (Task → specific detective)
  • /analyze command unchanged (launches codebase-detective)
  • Parallel orchestration patterns unchanged
  • All claudemem hooks preserved

Use Cases

When to Use Investigate SkillWhen to Use Direct Detective
Investigation type unclearYou know which specialist you need
General explorationParallel orchestration (multimodel plugin)
Quick routing decisionSpecific workflow requirements
Learning/experimentingProduction automation

Notes

  • Case-insensitive keyword matching
  • Priority system resolves conflicts
  • User sees routing decision before delegation
  • Original query preserved in Task context
  • Default to developer-detective when no keywords match
  • Works with all claudemem versions (v0.3.0+)

Maintained by: MadAppGang Plugin: code-analysis v3.1.0 Last Updated: January 2026 (v1.0.0 - Initial release)

Frequently asked questions

What does the Investigate AI skill do?

Unified entry point for code investigation. Auto-routes to specialized detective based on query keywords. Use when investigation type is unclear or for general exploration.

Why use Investigate on TypingMind?

Because you install it once and use it with any model. Investigate is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Investigate in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/MadAppGang/claude-code/tree/main/plugins/code-analysis/skills/investigate. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Investigate?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Investigate?

As many as you like. As long as a model supports skills, you can use Investigate with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Investigate AI skill free?

Yes. It is published on GitHub by MadAppGang under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇