eda-pdf-local-summary

スター8

フォーク3

更新日2026年3月6日 18:05

Summarize local paper PDFs into structured, citation-grounded evidence notes. Use when the user provides local PDF paths and asks for methods/assumptions/results/limitations extraction or wants evidence mapped to current EDA hypotheses.

インストール

Codex または Claude でインストールこの Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。

ソース

Mr-Fang-VLSI

Mr-Fang-VLSI/EDAgent

GitHub リポジトリを開く Creator のリポジトリを見る

ダウンロード

EDA PDF Local Summary

Overview

Use this skill to produce reusable local summary notes tied to page-level evidence.

Workflow

Validate input PDF path.

require absolute or repo-relative local path.
expected location: docs/papers/pdf/.

Extract text to local artifact.

Run scripts/extract_pdf_text.sh <pdf_path> <out_dir>.
default output dir: docs/papers/summaries/raw/.
default extractor priority is GROBID (via grobid_client + GROBID_SERVER), then falls back to pdftotext/mutool/pypdf if unavailable.
to disable GROBID for one run: set DISABLE_GROBID=1.
quality guard: if GROBID TEI result is too sparse, extractor auto-falls back to local text extraction.
when GROBID succeeds, keep TEI artifact alongside text (<pdf_stem>.tei.xml) for traceability.

Build structured summary.

Load references/summary-template.md.
Fill sections with explicit evidence snippets.
Prefer page-cited claims (p.xx) when recoverable from extracted text.
write output to: docs/papers/summaries/<pdf_stem>.summary.md.

3.5 Update paper manifest status.

locate corresponding row in docs/papers/manifests/*.paper_manifest.tsv.
update local_pdf_path and set status=summarized.

Sync landscapes after summary write.

run python3 scripts/common/paper_landscape_sync.py build.
ensure both landscape docs are updated:
- docs/knowledge_base/95_PROBLEM_LANDSCAPE.md
- docs/knowledge_base/96_METHOD_LANDSCAPE.md

Link to active hypothesis.

State whether paper supports, contradicts, or is neutral to current model/flow hypothesis.
add one "actionable next experiment" sentence.

Output Contract

Always output:

extracted text artifact path,
summary markdown path (docs/papers/summaries/<pdf_stem>.summary.md),
3-5 bullet findings with evidence strength (strong/medium/weak),
unresolved ambiguities requiring manual read,
updated manifest path (or explicit note if no manifest row matched).
landscape sync artifact paths (95/96 docs or explicit failure reason).

References

Load:

references/summary-template.md
references/evidence-grading.md

このリポジトリの他の Skills

同じリポジトリ

academic-presentation-crafter

Mr-Fang-VLSI/EDAgent

Creates and refines academic presentations from research papers. Use this skill for tasks involving the conversion of scientific or technical documents into slide decks for conferences, seminars, lectures, thesis defenses, or academic reviews. Supports beamer LaTeX (.tex), PowerPoint (.pptx), and Markdown output formats.

2026-04-288

delay-model-gate-evaluator

Mr-Fang-VLSI/EDAgent

Execute delay-model validation gates against HPWL baselines. Use when evaluating model consistency, running Gate-0 contract checks, building Gate-1 bucketed scorecards, performing wirelength-aware correlation checks, or deciding readiness for active optimization.

2026-04-288

workflow-research-chain

Mr-Fang-VLSI/EDAgent

Run a full EDA research chain from knowledge exploration and paper triage to idea debate, hypothesis experiment design, implementation, git versioning, validation, and retrospective. Use when the user requests an end-to-end EDA research workflow rather than a single isolated execution step.

2026-04-288

academic-writing-quality-assessor

Mr-Fang-VLSI/EDAgent

Audit and refine academic manuscripts for structure, terminology consistency, LaTeX hygiene, and high-signal grammar fixes. Use this skill for theses, papers, proposals, and technical reports when the task is manuscript writing rather than experiment execution.

2026-03-278

backside-routing-realization-specialist

Mr-Fang-VLSI/EDAgent

Diagnose and implement backside-routing realization paths, including targeted reroute, local DEF/OpenDB patching, and OpenROAD-backed net-level rerouter bring-up, when theory predicts benefit but the current flow does not realize BM2/BM1 usage.

2026-03-088

bspdn-physical-contract-auditor

Mr-Fang-VLSI/EDAgent

Audit whether the local BSPDN physical contract is coherent across paper assumptions, GT3 tech collateral, layer/via topology, and current flow policy before promotion or expensive attribution experiments.

2026-03-088

name	eda-pdf-local-summary
description	Summarize local paper PDFs into structured, citation-grounded evidence notes. Use when the user provides local PDF paths and asks for methods/assumptions/results/limitations extraction or wants evidence mapped to current EDA hypotheses.

EDA PDF Local Summary

Overview

Use this skill to produce reusable local summary notes tied to page-level evidence.

Workflow

Validate input PDF path.

require absolute or repo-relative local path.
expected location: docs/papers/pdf/.

Extract text to local artifact.

Run scripts/extract_pdf_text.sh <pdf_path> <out_dir>.
default output dir: docs/papers/summaries/raw/.
default extractor priority is GROBID (via grobid_client + GROBID_SERVER), then falls back to pdftotext/mutool/pypdf if unavailable.
to disable GROBID for one run: set DISABLE_GROBID=1.
quality guard: if GROBID TEI result is too sparse, extractor auto-falls back to local text extraction.
when GROBID succeeds, keep TEI artifact alongside text (<pdf_stem>.tei.xml) for traceability.

Build structured summary.

Load references/summary-template.md.
Fill sections with explicit evidence snippets.
Prefer page-cited claims (p.xx) when recoverable from extracted text.
write output to: docs/papers/summaries/<pdf_stem>.summary.md.

3.5 Update paper manifest status.

locate corresponding row in docs/papers/manifests/*.paper_manifest.tsv.
update local_pdf_path and set status=summarized.

Sync landscapes after summary write.

run python3 scripts/common/paper_landscape_sync.py build.
ensure both landscape docs are updated:
- docs/knowledge_base/95_PROBLEM_LANDSCAPE.md
- docs/knowledge_base/96_METHOD_LANDSCAPE.md

Link to active hypothesis.

State whether paper supports, contradicts, or is neutral to current model/flow hypothesis.
add one "actionable next experiment" sentence.

Output Contract

Always output:

extracted text artifact path,
summary markdown path (docs/papers/summaries/<pdf_stem>.summary.md),
3-5 bullet findings with evidence strength (strong/medium/weak),
unresolved ambiguities requiring manual read,
updated manifest path (or explicit note if no manifest row matched).
landscape sync artifact paths (95/96 docs or explicit failure reason).

References

Load:

references/summary-template.md
references/evidence-grading.md