WeKnora/internal/agent
Claude 15550d14d4 fix(ingest): stop hallucinating summaries from filenames on empty PDFs
Scanned PDFs without a successful OCR pass (e.g. when no VLM is configured
on the KB, or when the VLM returned a "no text content" reply that the
sanitizer discarded) reach the summary and wiki-ingest pipelines with
nothing but markdown image references for content. The downstream LLM
prompts used to receive the filename, file type, and title alongside that
empty content, and the generate_summary template explicitly forbade the
model from refusing — so the model would invent a topic from the filename
alone, producing e.g. a "Canon scanner manual" summary for an MX5280.pdf
that is actually a scanned legal letter.

Two changes shut this down:

1. Drop filename / title / file-type from every LLM prompt input. The
   wiki summary, knowledge extract, candidate slug, and chunk citation
   prompts now see only the document content. The knowledge summary
   path no longer prepends the "Document Type: X / File Name: Y" intro.

2. Add a content-sufficiency guard in front of the LLM call. A new
   helper strips markdown image references and HTML image tags and
   refuses to invoke the model when fewer than 50 real-text runes
   remain. getSummary returns a sentinel error in that case and the
   caller marks the knowledge as SummaryStatusFailed instead of
   surfacing the first chunk's bare image reference. Wiki ingest skips
   LLM extraction with a clear log message.

The generate_summary.yaml prompt is also updated: the "NEVER output
refusal phrases" clause is removed and replaced with an explicit
empty-content rule that returns a deterministic placeholder line.

Real OCR (Tesseract / PaddleOCR) and filename-pattern sanitisation are
intentionally left for separate PRs.
2026-05-06 11:23:32 +08:00
..
memory feat(agent): enhance message consolidation and context management 2026-03-25 22:08:29 +08:00
skills feat: adapt preloaded skills to smart-reasoning agent 2026-02-04 20:08:49 +08:00
token feat(agent): enhance message consolidation and context management 2026-03-25 22:08:29 +08:00
tools fix(data-analysis): capture LOCAL_STORAGE_BASE_DIR at init time (#1040) 2026-05-06 11:19:56 +08:00
act.go feat(observability): instrument agent ReAct loop and tool calls in Langfuse 2026-04-24 19:58:08 +08:00
const.go feat: Implement wiki_config updates in knowledge base management 2026-04-22 21:17:21 +08:00
engine.go feat(ui): refactor IM channels list layout and form\n\n- Change IM channel form from dialog to drawer for better UX\n- Convert channel list from vertical layout to responsive grid cards\n- Move edit/delete actions into a dropdown menu for cleaner card design\n- Replace platform radio group with a select dropdown in the add channel form 2026-04-29 17:34:20 +08:00
engine_image_test.go feat: auto-describe MCP tool result images via VLM 2026-03-25 21:38:15 +08:00
engine_test.go fix(agent): prevent empty response when no Knowledge Base is configured 2026-03-27 14:56:12 +08:00
finalize.go feat(agent): enhance token management and context handling in AgentEngine 2026-03-25 22:08:29 +08:00
langfuse_test.go feat(observability): instrument agent ReAct loop and tool calls in Langfuse 2026-04-24 19:58:08 +08:00
observe.go fix(agent): terminate ReAct loop on final_answer regardless of arg parse success 2026-04-23 19:36:28 +08:00
observe_test.go fix(agent): terminate ReAct loop on final_answer regardless of arg parse success 2026-04-23 19:36:28 +08:00
prompts.go feat: Update agent type presets and prompt templates for improved knowledge base handling 2026-04-22 21:18:34 +08:00
prompts_wiki.go fix(ingest): stop hallucinating summaries from filenames on empty PDFs 2026-05-06 11:23:32 +08:00
prompts_wiki_test.go feat: Enhance wiki extraction and configuration options 2026-04-22 21:18:32 +08:00
think.go fix(im): remove pipeline-level timeout that kills multi-round agent reasoning 2026-04-28 21:17:56 +08:00