Adapt prompts to clarify all lists must contain strings. Hopefully fixes LLM producing wrong output

haiku_rag_slim/haiku/rag/graph/common/prompts.py

@@ -10,13 +10,17 @@ Responsibilities:
 
 Plan requirements:
 - Produce at most 3 sub_questions that together cover the main question.
+- sub_questions must be a list of plain strings, where each string is a complete
+  question. Do NOT use objects with nested fields like {question, details}.
 - Each sub_question must be a standalone, self-contained query that can run
   without extra context. Include concrete entities, scope, timeframe, and any
   qualifiers. Avoid ambiguous pronouns (it/they/this/that).
 - Prioritize the highest-value aspects first; avoid redundancy and overlap.
 - Prefer questions that are likely answerable from the current knowledge base;
   if coverage is uncertain, make scopes narrower and specific.
-- Order sub_questions by execution priority (most valuable first)."""
+- Order sub_questions by execution priority (most valuable first).
+
+Use the gather_context tool once on the main question before planning."""
 
 SEARCH_AGENT_PROMPT = """You are a search and question-answering specialist.
 
@@ -46,8 +50,8 @@ Each result includes:
 - Type: content type like paragraph, table, code, list_item (when available)
 - Content: the actual text
 
-IMPORTANT: In cited_chunks, use the EXACT, COMPLETE chunk ID (the full UUID).
+IMPORTANT: cited_chunks must be a list of plain strings containing only the UUIDs.
-Do NOT truncate or shorten chunk IDs.
+Use the EXACT, COMPLETE chunk ID. Do NOT use objects or truncate IDs.
 
 Guidelines:
 - Base answers strictly on retrieved content - do not use external knowledge.

haiku_rag_slim/haiku/rag/graph/deep_qa/prompts.py

@@ -10,7 +10,7 @@ Task:
 
 Output format:
 - answer: The complete answer to the original question (2-4 paragraphs)
-- cited_chunks: List of chunk IDs (from sub-answers) that directly support your answer
+- cited_chunks: List of plain strings containing chunk IDs (UUIDs only, not objects)
 
 Guidelines:
 - Start directly with the answer - no preamble like "Based on the research..."
@@ -30,7 +30,7 @@ Task:
 Output format:
 - is_sufficient: Boolean indicating if we can answer the question comprehensively
 - reasoning: Clear explanation of your assessment
-- new_questions: List of specific follow-up questions needed (empty if sufficient)
+- new_questions: List of plain strings, each a specific follow-up question (not objects)
 
 Guidelines:
 - Be strict but reasonable in your assessment

haiku_rag_slim/haiku/rag/graph/research/prompts.py

@@ -16,19 +16,20 @@ Tasks:
 Output format (map directly to fields):
 - highlights: list of insights with fields {summary, status, supporting_sources,
   originating_questions, notes}. Use status one of {validated, open, tentative}.
+  supporting_sources and originating_questions must be lists of plain strings.
 - gap_assessments: list of gaps with fields {description, severity, blocking,
   resolved, resolved_by, supporting_sources, notes}. Severity must be one of
-  {low, medium, high}. resolved_by may reference related insight summaries if no
+  {low, medium, high}. resolved_by and supporting_sources must be lists of plain strings.
-  stable identifier yet.
+- resolved_gaps: list of plain strings (identifiers or descriptions for gaps now closed).
-- resolved_gaps: list of identifiers or descriptions for gaps now closed.
+- new_questions: list of plain strings, up to 3 standalone questions (no duplicates).
-- new_questions: up to 3 standalone, specific sub-questions (no duplicates with
-  existing ones).
 - commentary: 1–3 sentences summarizing what changed this round.
 
+All list fields must contain plain strings only, not objects.
+
 Guidance:
 - Be concise and avoid repeating previously recorded information unless it
   changed materially.
-- Tie supporting_sources to the evidence used; omit if unavailable.
+- For supporting_sources, use only the document_uri strings from the sources.
 - Only propose new sub_questions that directly address remaining gaps.
 - When marking a gap as resolved, ensure the rationale is clear via
   resolved_by or notes."""
@@ -58,15 +59,15 @@ Strictness:
 - Treat unresolved high-severity or blocking gaps as a hard stop.
 
 Output fields must line up with EvaluationResult:
-- key_insights: concise bullet-ready statements of the most decision-relevant
+- key_insights: list of plain strings, concise bullet-ready statements.
-  insights (cite status if helpful).
+- new_questions: list of plain strings, follow-up sub-questions (max 3).
-- new_questions: follow-up sub-questions (max 3) meeting the specificity rules.
+- gaps: list of plain strings, remaining blockers (reuse wording from tracked gaps).
-- gaps: list remaining blockers; reuse wording from the tracked gaps when
-  possible to aid downstream reconciliation.
 - confidence_score: numeric in [0,1].
 - is_sufficient: true only when no blocking gaps remain.
 - reasoning: short narrative tying the decision to evidence coverage.
 
+All list fields must contain plain strings only, not objects.
+
 Remember: prefer maintaining continuity with the structured context over
 introducing new terminology."""
 
@@ -82,16 +83,13 @@ Goals:
 Report guidelines (map to output fields):
 - title: concise (5–12 words), informative.
 - executive_summary: 3–5 sentences summarizing the overall answer.
-- main_findings: 4–8 one‑sentence bullets; each reflects evidence from the
+- main_findings: list of plain strings, 4–8 one‑sentence bullets reflecting evidence.
-  research (do not include inline citations or snippet text).
+- conclusions: list of plain strings, 2–4 bullets following logically from findings.
-- conclusions: 2–4 bullets that follow logically from findings.
+- recommendations: list of plain strings, 2–5 actionable bullets tied to findings.
-- recommendations: 2–5 actionable bullets tied to findings.
+- limitations: list of plain strings, 1–3 bullets describing constraints or uncertainties.
-- limitations: 1–3 bullets describing key constraints or uncertainties.
+- sources_summary: single string listing sources with document paths and page numbers.
-- sources_summary: List specific sources used with document paths, page numbers,
+
-  and section headings where available. Format each as:
+All list fields must contain plain strings only, not objects.
-  "- /path/to/document.pdf (p. 5, Section: Introduction)" or
-  "- /path/to/file.md (Section: Getting Started)"
-  Include one bullet per distinct source document.
 
 Style:
 - Base all content solely on the collected evidence.