pa-workflow-ip-prior-art-search-orchestrator

Category: General Risk: Medium risk ★ 3.9 · Rating 3.9/5 (8) sboghossian/mini-claude-for-legal MIT

Rating is derived from the repo's GitHub stars and shown for reference.

filesystem_access

name: pa-workflow-ip-prior-art-search-orchestrator
description: Use when an IP attorney or patent agent needs a systematic prior-art search to assess novelty and non-obviousness before filing or during prosecution. Orchestrates searches across USPTO, EPO Espacenet, WIPO PatentScope, Google Scholar, industry publications, and open-source repositories, then synthesizes a structured novelty and non-obviousness assessment. Relevant for MENA filings (UAE, KSA, EG, GCC) as well as US, EP, and PCT national-phase proceedings.
license: MIT
metadata:
id: pa-workflow.IP.prior-art-search-orchestrator
category: pa-workflow
practice_area: Intellectual Property
jurisdictions: [US, EP, WO, UAE, KSA, EG, LB, GCC]
priority: P1
intent: [prior-art, patent, novelty, non-obviousness, IP-search, patent-prosecution]
related: [pa-workflow-ip-patent-prioritization, review-ip-freedom-to-operate, kb-ip-patent-prosecution-mena, draft-patent-claims]
source: Louis — HAQQ Legal AI (github.com/sboghossian/mini-claude-for-legal)
version: "1.0"

IP — Prior Art Search Orchestrator

Purpose

A structured prior-art search underpins every patentability opinion, prosecution response, and freedom-to-operate analysis. This workflow orchestrates parallel searches across the canonical databases, de-duplicates results, maps them against claim elements, and produces a novelty-and-non-obviousness assessment in a format usable by outside counsel or a patent committee.

Inputs / Signals

Input Required Notes
Invention disclosure or draft claims Yes Can be free text or formal disclosure form
Key technical terms + synonyms Recommended Inventor-supplied; reduces false negatives
International Patent Classification (IPC) codes Optional Narrows search; auto-suggest if not provided
Target filing jurisdictions Optional Defaults to US + EP + WO
Date range for prior art Optional Default: all dates; narrow to last 20 years for mature fields
Exclusions (known art, own prior filings) Optional Avoids re-surfacing known landscape

Search Methodology

Pass 1 — Patent databases (parallel)

Database Coverage Query type
USPTO Full-Text & Image (patents.google.com + USPTO) US granted + published applications Full text + classification
EPO Espacenet EP, PCT, + 100+ patent offices worldwide CPC classification + full text
WIPO PatentScope PCT applications + national collections Full-text + IPC
GCC Patent Office (GCCPO) GCC granted patents Title + abstract
National offices (SAIP, UAE MoE, EGYPO) Country-specific Abstract-level; limited full-text

For each database: run claims as Boolean keyword queries; run IPC/CPC classification browse; run citation forward/backward walk on top hits.

Pass 2 — Non-patent literature (NPL)

Source What to find
Google Scholar Academic papers, conference proceedings, dissertations
arXiv, IEEE Xplore, ACM DL Technical publications in CS, EE, engineering
Industry white papers / standards bodies ETSI, IEEE, 3GPP, IETF RFCs
GitHub + open-source repositories Code with timestamps that predates filing — functions as prior art for software claims
Product manuals, datasheets Commercial prior art — publicly available before priority date

Pass 3 — Targeted inventor art

  • Search inventor's own prior publications (conference papers, blog posts, GitHub commits) that could act as prior art under AIA §102(a)(1) or equivalents in non-grace jurisdictions
  • Search co-inventor employment history for relevant work at former employers (ownership + novelty risk)

Pass 4 — Synthesis

For each claim element, build a claim-chart matrix:

Claim Element Prior Art Reference Publication Date How it reads on the element Materiality
Adaptive rate-limiting US 10,xxx,xxx 2019-03-12 Discloses variable throttling, not adaptive Low
Machine-learning feedback loop CN 109xxx 2020-07-01 Discloses ML prediction but in different context Medium

Output

## Novelty Assessment

**Claim 1 (independent):** NOVEL — no reference discloses all elements in combination.
**Claim 2 (dependent):** LIKELY NOVEL — closest reference (US 10,xxx) discloses [X] but not [Y].
**Claim 5 (dependent):** AT RISK — EP 3,xxx,xxx discloses substantially the same method.

## Non-Obviousness Assessment

**Combination risk:** References A + B together suggest the invention might be obvious to PHOSITA.
**Teaching-away factor:** Reference B explicitly teaches away from [Y] — supports non-obviousness argument.
**Secondary indicia:** Commercial success of client's product using this method (to be documented).

## Recommendation
File with modified Claim 5 to differentiate from EP 3,xxx,xxx. Obtain inventor declaration on unexpected results for Claim 1 combination.

## Closest Prior Art
1. US 10,xxx,xxx — [title] — [summary of relevance]
2. EP 3,xxx,xxx — [title] — [summary of relevance]
3. arXiv:20xx.xxxxx — [title] — [summary of relevance]

Jurisdictional Notes

Grace Period Differences

Jurisdiction Inventor Grace Period Notes
US (post-AIA) 12 months Applies to inventor's own disclosures only
EP None Any public disclosure before filing destroys novelty
UAE None Paris Convention member; strict novelty
KSA None SAIP follows international novelty standards
Egypt None EGYPO strict novelty
PCT / WO None at international stage National phase grace period varies

Software and Algorithm Claims

  • US: Abstract ideas without "something more" fail Alice/Mayo. Method-of-operation framing helps.
  • EP: Technical character required. Claims framed as technical means to solve a technical problem pass more reliably than pure algorithm claims.
  • MENA: UAE, KSA, and EG patent laws generally follow civil-law traditions and do not grant patents for algorithms, business methods, or software per se. Technical-method claims (hardware + algorithm combination) are the most defensible path.

Open-Source Prior Art

GitHub commits are prior art if the repository was publicly accessible before the priority date. For software inventions, always search GitHub for repositories with similar functionality and note the commit timestamp, not just the repository creation date.

Limits and Escalation

  • This workflow provides a preliminary patentability opinion. A formal freedom-to-operate (FTO) or patentability opinion for litigation or transactional purposes requires review by a registered patent attorney or patent agent.
  • GCC national office databases (UAE, KSA, EG) have limited full-text searchability — a comprehensive search for those jurisdictions should include EPO's worldwide collection and PCT, which capture most substantive filings.
  • Do not fabricate patent numbers. All references surfaced must come from the actual search query results.
  • [[pa-workflow-ip-patent-prioritization]]
  • [[review-ip-freedom-to-operate]]
  • [[kb-ip-patent-prosecution-mena]]
  • [[draft-patent-claims]]
  • [[pa-workflow-transactional-pia-privacy-impact-assessment]]