chasballew

TPRM Pipeline Prompt Engineering Improvements

For agentic workers: REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended) or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (- [ ]) syntax for tracking.

Goal: Apply Anthropic's official Claude 4.6 prompt engineering best practices to the assessment pipeline — enable adaptive thinking, restructure system/user prompt split, optimize long-context ordering, and clean up the prompt.

Architecture: The methodology prompt moves from the user message into the system prompt (enabling prompt caching). The user message is reordered to data-first/instructions-last. Adaptive thinking is enabled for the assessment stage via api.py. A new v4 prompt file replaces v3 with XML tags, trimmed content, and positive framing.

Tech Stack: Python 3.9+, Anthropic SDK 0.64.0, Claude Opus 4.6

chasballew

Prompt Engineering Review: vendor-risk-assessment-prompt-v3.md

Review of the TPRM pipeline's main assessment prompt against Anthropic's official prompt engineering best practices for Claude 4.6.

---

What's already strong

The prompt is well above average. The role definition is specific and purposeful. The citation standard with document register codes and evidence classes is excellent — it gives Claude concrete, unambiguous rules. The SOC 2 whitelist restriction includes a "why" explanation, which is exactly what the best practices recommend. The Four-Decision Process provides clear sequential gates. The style rules are sharp and actionable.

chasballew

Join the Aptible Gridiron GDPR Slack Community here.

Read the Aptible Privacy Statement to learn:

• What information we collect and why we collect it
• How we use that information and when we disclose it
• Your rights regarding that information, including how to access and update your information
• The steps we take to protect your information

Aptible Gridiron GDPR Slack Code of Conduct

chasballew

# http://nbviewer.ipython.org/url/norvig.com/ipython/Cheryl.ipynb

DATES = ['May 15',    'May 16',    'May 19',
         'June 17',   'June 18',
         'July 14',   'July 16',
         'August 14', 'August 15', 'August 17']

def month(date)
  date.split[0]
end

chasballew

# DOCKER-VERSION 0.4.0
from ubuntu:12.04
run echo 'deb http://us.archive.ubuntu.com/ubuntu/ precise universe' >> /etc/apt/sources.list
run apt-get -y update
run apt-get -y install python python-pip build-essential
run pip install flask flask-sqlalchemy
run apt-get -y purge build-essential
run apt-get -y autoremove
 
ADD . /opt/app

chasballew

# DOCKER-VERSION 0.4.0
from ubuntu:12.04
run echo 'deb http://us.archive.ubuntu.com/ubuntu/ precise universe' >> /etc/apt/sources.list
run apt-get -y update
run apt-get -y install python python-pip build-essential
run pip install flask flask-sqlalchemy
run apt-get -y purge build-essential
run apt-get -y autoremove

ADD . /opt/app

chasballew

Bash provides an environment variable called PROMPT_COMMAND. The contents of this variable are executed as a regular Bash command just before Bash displays a prompt[0]. You can use the output of regular commands in PROMPT_COMMAND. Here we use echo to construct the command to set the window title.

export PROMPT_COMMAND='echo -ne "\033]0;${PWD##*/}\007"'

The key sequence here is:
ESC]0;stringBEL -- Set icon name and window title to $string

Breaking it down:

chasballew

unless File.exist?('Gemfile')
  File.write('Gemfile', <<-GEMFILE)
    source 'https://rubygems.org'
    gem 'rails', github: 'rails/rails'
    gem 'arel', github: 'rails/arel'
    gem 'sqlite3'
  GEMFILE

  system 'bundle'
end