"No.

This is engagement-farm slop. The replies are more useful than the prompts — the commenters (Tom Martell, @ImmutableLawyer, @aniaacw, @OneManSaas) are all saying what you already know.

The prompts themselves are single-shot, zero-infrastructure, no-guardrail templates with these specific failures:

Persona inflation is the whole trick, and it’s the wrong trick. “You are a senior corporate attorney at Skadden Arps” does nothing useful. The firm name doesn’t change the model’s legal knowledge — it just inflates confidence and suppresses hedging, which is the opposite of what you want for legal output. Your Gem suite uses grounded role framing (“You draft discovery requests for a California PI plaintiff firm”) tied to actual jurisdiction, practice area, and output constraints. That’s the difference between a persona that anchors behavior and one that just sounds impressive.

No verification layer, no placeholder protocol, no QC. Every one of these prompts asks for “ready-to-sign” output. Your entire architecture is built around the opposite principle — [NEED] markers, [CITE CHECK] placeholders, mandatory Gem 6 QC pass, human verification of all arithmetic. These prompts actively encourage the user to skip the step that matters most.

No jurisdiction pinning. The prompts say things like “state-specific enforceability” for non-competes but never force the model to ask which state. California bans most non-competes entirely (Bus. & Prof. Code § 16600). A “ready-to-sign” employment agreement with a non-compete clause drafted for a California employee is malpractice-grade output.

No hallucination controls. No citation verification, no confidence flagging, no instruction to say “I don’t know” when the model is uncertain about a legal standard. These prompts are optimized to produce confident-sounding output, not correct output.

Zero platform optimization. Every prompt is the same wall-of-text format regardless of target model. No thinking budget management, no verbosity control, no structural formatting matched to the platform.

The one commenter who nailed it: Tom Martell — “Claude is the best junior associate I’ve ever had but you don’t have to sensationalize the use case.” That’s the right frame. Your infrastructure already treats AI as a junior associate with mandatory supervision. These prompts treat it as a replacement for the supervising attorney. Nothing to extract here.​​​​​​​​​​​​​​​​

(http://www.autoadmit.com/thread.php?thread_id=5847693&forum_id=2...id.#49756514)