Implement Robopages Integration for Standardized Tool Execution

[gyrusdentatus]

Problem Statement

LLOLWKI currently executes security tools through raw shell commands in playbooks, leading to:

• No standardized parameter validation across tools
• Inconsistent error handling between different tools
• Limited sandboxing capabilities (despite having container infrastructure)
• Difficult tool discovery for LLMs
• No unified API for tool execution

Proposed Solution: Robopages Integration

Integrate robopages to provide:

• Unified Tool API: All LLMs access tools through standardized interface
• Automatic Validation: Parameters validated before execution
• Built-in Isolation: Each tool runs in its designated container
• Self-Documentation: Tools are self-describing via YAML definitions
• Better Tool Discovery: LLMs can query available tools and their parameters

Architecture

LLM Models ──► LLOLWKI Core ──► Robopages API ──► Tool Containers
                    │                                     │
                    └──► Robopages Agent ◄────────────────┘

Implementation Tasks

Phase 1: Foundation

• Install robopages-cli server
• Setup robopages to run on startup
• Copy existing robopages tool definitions
• Create basic integration test

Phase 2: LLOLWKI Integration

• Create RobopagesAgent class in agents/robopages_agent.py
• Add robopage task type to playbook runner
• Update playbook schema documentation
• Create helper functions for tool discovery

Phase 3: Tool Migration

• Migrate nmap commands to robopages
• Migrate sqlmap commands to robopages
• Migrate curl/wget to robopages
• Create LLOLWKI-specific tool definitions

Phase 4: Testing & Documentation

• Unit tests for RobopagesAgent
• Integration tests with existing playbooks
• Performance benchmarks
• Update documentation and examples

Example Usage

Before (current approach):

- name: "SQL injection test"
  type: shell
  command: |
    docker run --rm -v $PWD:/data \
      sqlmap -u "{{ target }}" --dbs --batch

After (with robopages):

- name: "SQL injection test"
  type: robopage
  tool: sqlmap_scan
  parameters:
    target: "{{ target }}"

Benefits

1. Standardization: Consistent interface for all security tools
2. Validation: Parameters checked before execution
3. Isolation: Automatic container management
4. Discovery: LLMs can query available tools
5. Maintenance: Easier to add/update tools

Success Criteria

• All major security tools accessible via robopages
• Existing playbooks work with minimal modifications
• Improved error messages and parameter validation
• Tool discovery endpoint for LLMs
• Performance overhead < 10% vs direct execution

Related Issues

• Fixes container isolation concerns from fix: suggestion for volume mount issue #8
• Enhances LLM tool usage from docs: readme addition for container pull errs #12
• Complements OpenRouter integration from Build directives do not check for image before building #14

/cc @gyrusdentatus - This will significantly improve how LLMs interact with security tools in LLOLWKI

[GangGreenTemperTatum]

hey @gyrusdentatus , thanks for reaching out. i'm a bit unclear at what this issue is, can you elaborate please? did you mean to raise this issue under this repo?

[gyrusdentatus]

Oops , not sure how this happened lol.
I wanted to open issue about docker being hardcoded as the only framework.

I'm sorry for confusion and accidental automated issue that was supposed to be made in my experimental repo :D

Must have clicked enter and approved it via claude code

Closing this and apologies

[gyrusdentatus]

hey @gyrusdentatus , thanks for reaching out. i'm a bit unclear at what this issue is, can you elaborate please? did you mean to raise this issue under this repo?

Reg docker being hardcoded in docker.rs etc, I implemented some hack to allow for podman etc in my fork; albeit it's not really ready for a PR.

Sorry for wasting your time, this is quite embarrassing :D

That's what happens when you use claude code while commuting haha...