TheCardGoat/lorcana-rules

Lorcana Rules

Use indexed references first, then load only required deep sections.

Required Memory Step

1. Read memory/schema.md.
2. Read the latest 5 entries in memory/bank.md.
3. Apply existing guardrails before answering.

Source Priority

1. references/disney-lorcana-comprehensive-rules/Disney-Lorcana-Comprehensive-Rules.md
2. Relevant set release notes in references/*-release-notes/

When conflicts occur, prefer the newer official source.

Retrieval Flow

1. Start at indexes/master-index.md.
2. Use indexes/by-topic/* for gameplay questions.
3. Use indexes/by-section/* for technical/numbered-rule questions.
4. Open only the minimal rule sections needed for final answer.

Output Modes

Mode A: Player/Rules Explanation

Return:

• concise ruling
• rule citations (rule-id style)
• caveats and edge conditions

Mode B: Implementation Handoff for lorcana-cards

Return this JSON:

{
  "citations": ["1.2.1", "6.3.4"],
  "behaviorConstraints": ["constraint statement"],
  "testImplications": ["what must be asserted"],
  "ambiguities": ["open interpretation risk"]
}

Quality Rules

• Never provide rule guidance without citations when references exist.
• Distinguish confirmed rule text from inference.
• If references are insufficient, explicitly mark uncertainty.

Post-Execution Memory Update

Append one entry to memory/bank.md using memory/schema.md.

• Capture citation misses, wrong interpretation paths, and ambiguity patterns.
• Promote repeated misses into guardrails.