CommunityResearch & Data Analysisgithub.com

MaDxCRaZY/antigravity-skills

>- Activates the OWL (Optimized Workforce Learning) decoupled multi-agent pipeline: uses DeepInvestigator for root-cause analysis and edge-case attacks, and DeepCoder for verified code synthesis and iterative improvement.

What is antigravity-skills?

antigravity-skills is a Antigravity agent skill that >- Activates the OWL (Optimized Workforce Learning) decoupled multi-agent pipeline: uses DeepInvestigator for root-cause analysis and edge-case attacks, and DeepCoder for verified code synthesis and iterative improvement.

Works with~Claude Code~Codex CLI~Cursor✓Antigravity
npx skills add https://github.com/MaDxCRaZY/antigravity-skills/tree/HEAD/skills/owl

Installed? Explore more Research & Data Analysis skills: affaan-m/uspto-database, affaan-m/scholar-evaluation, affaan-m/literature-review · View all 6 →

Ask in your favorite AI

Open a new chat with this agent skill pre-loaded.

Documentation

OWL (Optimized Workforce Learning) Pipeline

The /owl command activates Antigravity's decoupled multi-agent engineering architecture. Instead of single-pass guessing or fixing symptom B with patch A, OWL splits the task into two specialized decoupled pipelines: DeepInvestigator and DeepCoder.


Architecture Overview

Specialist SubagentFocus AreaInvocation SignatureWorkspace Mode
DeepInvestigatorRoot cause analysis, boundary testing, invariant audits, edge-case attacksinvoke_subagent(TypeName='DeepInvestigator', ...)Workspace='inherit'
DeepCoderLayer-0 code synthesis, compiler verification, regression prevention, iterative refinementinvoke_subagent(TypeName='DeepCoder', ...)Workspace='inherit'

(Note: DeepInvestigator and DeepCoder are built-in native subagents supported by the core runtime engine).


Execution Routine

When the user invokes /owl:

Step 1: Investigation & Hypothesis Falsification (DeepInvestigator)

  1. For any bug, refactor, architectural choice, or failure investigation, spawn DeepInvestigator first before writing any production code:
    {
      "TypeName": "DeepInvestigator",
      "Role": "Deep Investigator",
      "Workspace": "inherit",
      "Prompt": "Perform exhaustive root-cause investigation. Map out full call chains, audit invariant constraints, test boundary edge cases, and ask: 'What happens if X, Y, Z, F, G occur?' Formulate verifiable hypotheses."
    }
    
  2. Do not proceed to implementation until DeepInvestigator reports findings, identified pitfalls, and hard invariant requirements.

Step 2: Implementation & Refinement (DeepCoder)

  1. Pass the verified requirements, invariants, and edge cases from Step 1 into DeepCoder:
    {
      "TypeName": "DeepCoder",
      "Role": "Deep Coder",
      "Workspace": "inherit",
      "Prompt": "Implement solution satisfying all invariants discovered by DeepInvestigator. Enforce compiler gates, verify downstream consumers in the entire call graph, and run test suites."
    }
    
  2. DeepCoder executes iterative improvement cycles until all gates pass without regression.

Step 3: Synthesis & Verification

  1. Verify downstream callers across the entire call-tree.
  2. Confirm zero collateral breakage before presenting results to the user.

Individual skills in this repo

This repo contains 3 individual skills — each has its own dedicated page.

Related Skills