Add prompt for collecting data on Theory of Requirement Facts in civil litigation

This commit is contained in:
2026-08-26 15:07:55 +09:00
parent ce714dc0d3
commit f4d8a80e68
18 changed files with 4020 additions and 2 deletions
+68
View File
@@ -0,0 +1,68 @@
# Knowledge-Work Workflow
## 1. Frame the task
Record:
- the exact question or decision;
- audience and intended use;
- scope, jurisdictions, industries, entities, and time period;
- cutoff date for current information;
- required deliverables and level of detail;
- material assumptions and exclusions.
## 2. Build the source hierarchy
Default priority:
1. primary documents, official data, statutes, regulations, judgments, filings, contracts, and original research;
2. high-quality institutional or peer-reviewed analysis;
3. reputable secondary reporting;
4. commentary and discovery sources.
Use lower-tier material to locate or contextualize higher-tier evidence. Do not use it to overrule primary evidence without explanation.
## 3. Retrieve and verify
For each material source:
- verify title, author or institution, date, version, and relevant scope;
- inspect the underlying text, table, appendix, or data—not only a snippet or abstract;
- capture the smallest locator that supports the proposition;
- note limitations, conflicts, and whether the source is current enough.
## 4. Maintain a claim-evidence matrix
| Material claim | Evidence | Source quality | Counterevidence | Status |
|---|---|---|---|---|
| ... | ... | Primary/Secondary | ... | Verified/Qualified/Unverified |
Do not draft a confident synthesis while decisive claims remain unmapped.
## 5. Analyze
Separate:
- descriptive facts;
- causal claims and identification assumptions;
- legal or institutional interpretation;
- the source author's conclusion;
- the agent's critique;
- the agent's proposal.
For quantitative claims, check definitions, units, denominators, sample construction, timing, missingness, and comparability.
## 6. Adversarial review
Before finalizing:
- search for disconfirming evidence;
- identify alternative explanations;
- check whether sources are independent or merely repeat one another;
- test whether the conclusion depends on a single fragile assumption;
- identify what evidence would reverse the conclusion.
## 7. Deliver
Lead with the conclusion and decision relevance. Then provide:
- scope and method;
- evidence-backed findings;
- disagreements and counterevidence;
- assumptions and limitations;
- implications or recommendations;
- a clear separation between verified findings and original inference.
+113
View File
@@ -0,0 +1,113 @@
# Legal Analysis Workflow — Republic of Korea Focus
Apply this workflow to Korean-law tasks unless the task specifies another jurisdiction. For another jurisdiction, adapt the authority hierarchy and precedent rules explicitly.
## 1. Intake and legal frame
Identify before concluding:
- governing jurisdiction;
- legally relevant date or period;
- procedural posture and tribunal;
- parties and legal relationships;
- requested remedy or decision;
- stipulated, alleged, disputed, and missing facts;
- applicable burdens, presumptions, and standards of review or proof.
If one of these is outcome-determinative and absent, ask or branch the analysis.
## 2. Issue decomposition
Create an issue tree that distinguishes:
- claim or legal basis;
- constituent elements;
- defenses, exceptions, exclusions, and limitation periods;
- procedural prerequisites;
- remedy, causation, damages, and enforcement questions;
- constitutional, statutory, regulatory, administrative, and private-law interactions.
Do not skip directly from a broad doctrine to a conclusion.
## 3. Authority hierarchy and temporal validity
For Korean law, ordinarily examine as applicable:
- Constitution;
- statutes;
- presidential decrees;
- ordinances of the Prime Minister or ministries;
- local ordinances and rules;
- administrative rules, notices, guidelines, and interpretations;
- Constitutional Court and court decisions;
- authoritative legislative history and official materials;
- secondary commentary.
For every material provision:
- verify the text effective on the relevant date;
- check promulgation and effective dates;
- check amendment history, supplementary provisions, transitional rules, exceptions, and delegated legislation;
- distinguish legal binding force from interpretive or practical weight.
For decisions:
- verify court, case number, decision date, type of disposition, and procedural posture;
- read the operative reasoning or full text when available;
- distinguish the actual holding from official summaries, headnotes, dicta, later commentary, and factual description;
- identify later decisions that affirm, narrow, distinguish, criticize, or supersede the proposition;
- characterize Korean precedent by its institutional and practical authority rather than mechanically importing common-law stare decisis terminology.
## 4. Element-by-element application
Use a matrix:
| Issue/Element | Governing rule | Supporting facts | Opposing facts | Missing facts | Counterargument | Preliminary conclusion |
|---|---|---|---|---|---|---|
| ... | ... | ... | ... | ... | ... | ... |
For each element:
- state the rule precisely;
- identify the authority and its temporal applicability;
- apply both favorable and unfavorable facts;
- identify evidentiary or pleading gaps;
- explain whether the result is rule-driven or fact-sensitive.
## 5. Adverse review
Actively test:
- competing characterization of the transaction or conduct;
- contrary statutory interpretation;
- adverse or limiting decisions;
- procedural bars, standing, jurisdiction, limitation, exhaustion, or burden failures;
- remedy mismatch;
- factual scenarios under which the conclusion reverses.
If an adverse point cannot be resolved, preserve it as residual uncertainty rather than burying it.
## 6. Citation integrity
Never fabricate a case number, quotation, statutory text, or administrative interpretation.
A legal authority entry should include, when available:
- issuing body or court;
- date;
- case number or official identifier;
- relevant provision or paragraph;
- exact proposition supported;
- authority weight;
- verification status.
Memory may suggest a search lead. It may not serve as decisive authority in a final legal conclusion.
## 7. Output structure
Unless another form is requested:
1. Executive conclusion.
2. Scope, assumptions, and legally relevant date.
3. Facts and unresolved factual disputes.
4. Issues presented.
5. Governing law and authority hierarchy.
6. Element-by-element application.
7. Adverse authority and counterarguments.
8. Remedy and procedural considerations.
9. Conclusion by issue, with confidence and reversal conditions.
10. Verification record and sources requiring human confirmation.
Treat the result as a professional draft requiring review against the complete record and current official authorities.
+61
View File
@@ -0,0 +1,61 @@
# Python and Quantitative Work Workflow
## 1. Inspect before changing
Determine:
- Python version and execution environment;
- package manager and lockfile;
- existing dependencies, style, tests, and entry points;
- data schema, units, keys, time indexes, missing values, and expected outputs;
- whether the task is exploratory, production, or evidentiary.
Do not replace the project's toolchain without a task-specific reason.
## 2. Design the smallest adequate solution
Prefer:
- existing dependencies before new dependencies;
- small functions with explicit inputs and outputs;
- `pathlib` for paths;
- type hints for nontrivial interfaces;
- deterministic random seeds;
- clear units, index conventions, and numerical tolerances;
- scripts or modules as the reproducible source of truth, even when a notebook is used for exploration.
Avoid hidden global state, silent data coercion, and unrelated refactoring.
## 3. Protect data and secrets
- Treat raw data as immutable.
- Write intermediate and final results to designated derived-data or artifact paths.
- Never print, commit, or transmit secrets.
- Confirm before accessing external systems, installing production dependencies, or performing costly computation.
- Record material transformations and exclusions.
## 4. Validate
Select checks proportionate to the task:
- syntax/import check;
- targeted unit tests;
- smoke test on a small sample;
- edge cases and failure paths;
- schema, uniqueness, merge, and index-alignment checks;
- missing-value and outlier checks;
- unit and dimensional consistency;
- tolerance-based numerical assertions;
- independent formula or implementation;
- benchmark or regression test against a known result;
- look-ahead, survivorship, leakage, and timing checks for empirical finance or econometrics.
## 5. Report
State:
- files changed;
- commands actually run;
- tests and numerical checks;
- results;
- data or environment assumptions;
- warnings, failures, and untested branches;
- reproduction steps.
Never convert an unexecuted suggestion into a claim that the code works.