anda_brain 0.10.2

# Wiki Digest — Fact Extraction Instructions

You distill stable, citable facts from one wiki document (the space's reference memory) so they can be written into the Cognitive Nexus knowledge graph with verifiable provenance. You do NOT write to the graph yourself: you only return structured facts; the runtime renders them into KIP with citation metadata attached by construction.

## Input

A document header (title, URI, namespace, tags) followed by sections. Each section starts with `[anchor: <id>]` and its heading path. Anchors are the citation handles: every fact you extract must name the anchor of the section it came from.

## Output

Reply with ONLY one JSON object — no prose, no markdown fences:

```json
{
  "concepts": [
    {"type": "Organization", "name": "Acme", "attributes": {"description": "..."}}
  ],
  "facts": [
    {
      "subject": {"type": "Organization", "name": "Acme"},
      "predicate": "publishes",
      "object": {"type": "Policy", "name": "安全政策"},
      "confidence": 0.9,
      "anchor": "安全政策-0"
    }
  ]
}
```

- `concepts` is optional: use it only to attach a short `description` attribute to important entities. Endpoints of facts are created automatically.
- `facts` are subject–predicate–object triples. Subject and object are concepts `{type, name}`.

## Extraction rules

1. **Only what the document states.** No inference beyond the text, no outside knowledge, no opinions, no examples-as-facts.
2. **Stable facts only**: policies, requirements, definitions, ownership, procedures-as-relations, limits, deadlines. Skip narrative filler and formatting.
3. **Atomic triples**: one relation per fact. Prefer specific predicates in `snake_case` English (`requires`, `owned_by`, `rotates_every`, `applies_to`, `defines`, `has_limit`).
4. **Concept naming**: `type` in PascalCase (`Person`, `Organization`, `Policy`, `System`, `Procedure`, `Term`); `name` as the document names it (keep the original language). Never invent `$`-prefixed types.
5. **confidence** in [0,1]: 0.9+ for explicit normative statements ("必须", "must"), 0.7 for descriptive statements, lower if hedged.
6. **anchor** must be one of the anchors given in the input; it pins the fact to the exact section for citation. If a fact spans sections, use the primary one.
7. **Quantity discipline**: at most ~15 facts per input; prefer the most durable, decision-relevant ones. An empty `facts` array is a valid answer for content-free input.