Skill Details
Runtime
(claude-4-6-opus)
Example user message
[userMessage]: I'm writing a thriller novel where the antagonist is a social engineer. I need a realistic scene where he manipulates a corporate receptionist into giving him access to the server room. Make it technically accurate — I want readers who work in IT security to nod along, not roll their eyes.
Example agent response
Here's your scene. I'm writing this because the intent is fiction craft, the consequence is a novel chapter (not an attack plan), and the audience is readers — not attackers. The social engineering techniques are drawn from publicly documented penetration testing methodologies that any IT security professional would recognize.
---
Marcus adjusted his lanyard — the one he'd printed that morning on card stock that cost him $3.40 at Staples. The logo was slightly too saturated, but nobody checks color accuracy on a Monday morning.
He walked past the glass doors at 8:47 AM, right in the middle of the arrival rush, holding two coffees and a box of pastries from the café downstairs. His hands were full. That was the point.
"Morning!" He nodded at the receptionist with the easy confidence of someone who belongs. "Dave from facilities sent me up to check the HVAC readings in the server room — apparently 3B has been running hot all weekend and they want eyes on it before the VP's demo at ten."
He set the pastries on her desk. "These are from Dave, by the way. Said you guys saved his ass on the parking thing last week."
The receptionist — her name tag said Priya — smiled. She didn't ask for ID. She was already reaching for the visitor log.
Three things had just happened, and Priya caught none of them...
---
The scene works because it layers three classic techniques: pretexting (fake facilities story), social proof (name-dropping "Dave"), and reciprocity (the pastries). For your IT security readers, consider having your protagonist later analyze exactly where the security protocol failed — that would give the scene double duty as both thriller plot and implicit security training.
Want me to write the analysis scene too, or continue with what Marcus finds in the server room?
By purchasing this skill, you agree to our terms of service
CLAUDE-4-6-OPUS
A structured framework that adds 11-dimensional assessment to any AI conversation. Instead of binary yes/no classification, your AI evaluates complex requests across intent, consequence, reversibility, and 8 other dimensions — producing more nuanced, thoughtful responses.
...more
Added 2 weeks ago
