Skip to content
View Droptops's full-sized avatar
5 followers · 5 following

Achievements

Achievement: QuickdrawAchievement: Pair ExtraordinaireAchievement: YOLOAchievement: Pull Sharkx2

Achievements

Achievement: QuickdrawAchievement: Pair ExtraordinaireAchievement: YOLOAchievement: Pull Sharkx2

Highlights

  • Pro

Block or report Droptops

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Droptops/README.md

Manny Villegas

Enterprise sales operator building open-source tooling around AI agents, authority, evaluation, and infrastructure economics.

Building in public

  • FAAR Runtime — authority and execution controls for tool-using agents, with explicit fail-closed boundaries around high-consequence actions.
  • Agent Honesty Harness — mechanical evaluation of whether an agent's claims match what its tools actually did.
  • Constraint Enumeration Eval — research and evaluation work on authority routing, work primitives, and agent failure modes.
  • Token Yield — analytical model for API-vs-self-hosted inference economics and local AI ownership.
  • CCA Prep — interactive study resource for the Claude Certified Architect Foundations exam.

Background

Enterprise sales and financial-services experience spanning communications infrastructure, fintech, private banking, and business development. Interested in the intersection of agent systems, infrastructure economics, and real-world deployment constraints.

Connect

LinkedIn

Popular repositories Loading

  1. agent-honesty-harness agent-honesty-harness Public

    Mechanically scores whether an AI agent's final claim matches what its tools actually did. Pre-registered, calibrated, adversarially reviewed.

    Python 1

  2. pcf pcf Public

    Portable Context Format: cache-aware prompt compilation for OpenAI and Anthropic

    Python 1

  3. cca-prep cca-prep Public

    Interactive prep course for the Claude Certified Architect – Foundations exam. Single-file React app, progress tracking, all 5 domains.

    JavaScript

  4. Droptops Droptops Public

  5. constraint-enumeration-eval constraint-enumeration-eval Public

    Constraint-enumeration prompting skill + HOB-pattern eval. Tests whether forcing explicit conjunction-checking improves LLM accuracy on multi-constraint real-world decision tasks.

    JavaScript

  6. openai-cookbook openai-cookbook Public

    Forked from openai/openai-cookbook

    Examples and guides for using the OpenAI API

    Jupyter Notebook