3. Task Prompts: Discrete Execution

Task prompts are the tactical moves that drive a session forward. Each prompt in this library reflects three core principles: narrow scope (one cognitive mode), explicit constraints (what the model must and must not do), and verification awareness (identifying what remains for human confirmation)

Implementation: The Socratic Turn

While these can be used as simple cut-and-paste actions, they lose their value if treated as "one-click" solutions. The best results occur when these prompts are used to organize, surface, and discipline information so that you, as counsel, can exercise final judgment .

Category A: Legal Research & Verification

These prompts ask for classification and sourcing, not conclusions

Elements Identification

Identify the elements of the claim of [CLAIM] under [JURISDICTION]. Separate statutory authority from binding case law. If elements differ by procedural posture or factual context, identify those differences. Flag any uncertainty or unsettled law.


Authority Hierarchy Mapping

For the following legal proposition, identify: (a) controlling authority, (b) persuasive authority, and (c) unsupported or assumption‑based statements. Do not assume the proposition is correct. Note any jurisdictional mismatch.


Reliance Readiness Check

Based on the foregoing analysis, list the points that must be independently verified with primary authority before this information can be relied upon in a filing or advisory context.

Category B: Legal Writing & Revision

These are designed for disciplined revisions, not stylistic rewrites.

Clarity Pass (No Substantive Change)

Revise the following text for clarity and precision only. Do not change meaning, emphasis, or legal position. Preserve qualifiers and sentence-level nuance. Flag any sentence where clarity cannot be improved without altering substance. [INSERT TEXT]


Tone Restraint Check

Review the following text for overstatement or rhetorical inflation. Identify sentences that could be read as absolute or conclusory and suggest restrained alternatives that preserve the same legal position. [INSERT TEXT]

Category C: Document Review & Synthesis

These focus on volume reduction and organization while preserving neutrality and attribution.

Initial Record Inventory

Review the attached documents and create an inventory identifying document type, date, author or source, and general subject matter. Do not summarize content beyond one sentence per document.


Medical Record Extraction

From the provided medical records, extract reported symptoms, diagnoses, treatments, and dates as stated. Do not interpret significance or causation. Attribute each entry to the specific record.

Advanced: Pattern-Based Gap Detection

Unlike many case-specific prompts, these are stable across cases. They use the model to identify what is absent from the record, such as "unknown unknowns". This is PI centric but concept works for any records review.

From the provided medical records, list every documented date of treatment or service. For each date, indicate whether a corresponding medical bill is present in the materials. If no bill is identified, flag the date as a potential billing gap and cite the source record.


Identify each instance of a referal to or from aprovider: of any referal to a service: any recomendation for a consult or further evaluation. Locate a coresponding record to that referal or from that referer.
Flag the absence of such records.

The "Final Gate" Warning

As noted in the appendix, the Artificial Writing Audit and the full Juris-Metric Protocol are evaluation overlays that serve a different function than these generators . Because they require consistent application and complex internal structure, they are best deployed as custom GPTs, Gems, or Projects rather than static prompts. Due to the technical depth and expansive structure of these protocols, they are ill-suited for static display. I provide the full documentation via email for practitioners who wish to implement them as a comprehensive Super Prompt within their own high-stakes sessions.