# Documentation Index

Use this index to select the smallest relevant workflow.

| Document | Use when |
|---|---|
| `PLANS.md` | The task is long, multi-stage, high-stakes, or requires several coordinated deliverables. |
| `workflows/knowledge-work.md` | Research, synthesis, academic critique, policy reports, document analysis, or evidence-backed recommendations. |
| `workflows/legal-analysis-ko.md` | Legal issue identification, authority research, statutory interpretation, case analysis, application, remedies, or litigation-oriented critique. |
| `workflows/python-work.md` | Python coding, data cleaning, quantitative checks, simulations, econometrics, or artifact generation. |
| `quality/evidence-policy.md` | The output relies on external sources, quotations, legal authorities, numerical evidence, or current facts. |
| `../templates/TASK_PROMPT.md` | A new task needs a precise goal/context/constraints/done specification. |
| `../evals/EVAL_RUBRIC.md` | Testing or revising `AGENTS.md`, workflows, model settings, or subagent design. |

Do not read every document by default. Read the documents required by the current task.