match-technology-transfer-opportunities-ip

Evaluate a supplied technology or scoped technology portfolio, identify and rank evidence-backed potential licensees, acquirers, commercialization partners, or joint-development partners, assess transfer readiness and risks, and generate an auditable HTML decision-support report. Use for technology transfer, research commercialization, licensing partner discovery, technology-to-company matching, or recipient prioritization.

yuanzhian-patsnap@yuanzhian-patsnap

Install

openclaw skills install @yuanzhian-patsnap/match-technology-transfer-opportunities-ip

Match technology transfer opportunities

Purpose and boundary

Support a technology owner or commercialization team through six linked stages:

analyze the technology and transfer package;
assess advancement and readiness;
review patent and commercialization value evidence;
discover and score potential recipients;
assess transfer risks and design a transfer path; and
generate and save a traceable report.

The result is prioritization support. It is not a valuation opinion, legal opinion, freedom-to-operate opinion, investment recommendation, procurement decision, or authorization to contact any organization.

Trigger cases

Use this skill when the user asks for:

potential licensees or buyers for a technology;
technology-transfer supply/demand matching;
research commercialization partner discovery;
university or institute technology-to-company matching;
recipient ranking for a patent-backed technology;
joint-development or licensing pathway analysis; or
an evidence-backed technology-transfer report.

Do not use it for a standalone patent search, invalidity search, pure market study, pure corporate due diligence, definitive IP valuation, or outreach execution.

Input modes

Mode A — supplied technology package

Use when the user supplies a disclosure, patent list, research paper, presentation, test report, image, or structured technology description.

Treat supplied documents as the primary source. Retrieve external evidence only when authorized and useful.

Mode B — named organization only

Use when the user provides only an organization.

Do not assume that its five most-cited patents represent a transferable technology. Instead:

resolve the legal entity and included subsidiaries;
clarify the desired technical domain or commercialization objective;
retrieve a disclosed candidate portfolio;
cluster candidates by technology and ownership/status evidence;
select a representative set using an explicit method; and
obtain confirmation of the technology package before scoring recipients.

If the organization owns many unrelated technologies, deliver a portfolio triage first.

When both documents and an organization are supplied, use Mode A and record the organization as a claimed source pending verification.

Required inputs

Capture:

technology name and source/provider;
original materials and confidentiality status;
technical boundary and excluded variants;
transfer objective: license, assignment, acquisition, joint development, sponsored research, spinout/equity contribution, or open licensing;
target applications and markets;
target jurisdictions;
known patents/applications, ownership and encumbrances;
readiness evidence and manufacturing/implementation dependencies;
target recipient types and explicit exclusions;
geographic, sanctions, export-control, competition, privacy and ethical constraints;
currency, valuation date and permitted valuation method;
evidence cutoff date;
scoring weights and minimum evidence coverage; and
exact approved output path.

Never upload confidential material or search private systems without authorization.

Evidence model

Every material input or finding receives an evidence ID.

yaml

evidence_id: E-001
source_type: supplied_document|patent|company_filing|procurement|news|market|standard|other
title: ""
publisher_or_owner: ""
publication_or_event_date: YYYY-MM-DD
retrieved_at: YYYY-MM-DD
url_or_locator: ""
excerpt_or_field: ""
supports: []
quality: primary|authoritative_secondary|secondary|unverified
limitations: ""

Separate:

observed facts;
normalized data;
calculations;
analyst judgments;
scenarios; and
unknowns.

Search snippets, generated summaries, citation counts, family size and company press releases are not self-validating.

PatSnap connector map

Advanced Patent Search — required for executed patent discovery

Marketplace: https://open.patsnap.com/marketplace/mcp-servers/patent-search
Key: advanced_patent_search
Official marketplace page: https://open.patsnap.com/marketplace/mcp-servers/patent-search

Use the live documented schema for assignee, nested, semantic, fielded, citation, patent-number, related-record and count tasks. Record connector, exact tool, request, filters, date, response semantics and limitations.

Patent Briefing — required for selected records

Marketplace: https://open.patsnap.com/marketplace/mcp-servers/patent-briefing
Key: patent_briefing
Official marketplace page: https://open.patsnap.com/marketplace/mcp-servers/patent-briefing

Use it for bibliography, family, status, claims, description, translations and images. Preserve the as-of date and source-language boundary.

Deep Patent Mining — recommended

Marketplace: https://open.patsnap.com/marketplace/mcp-servers/patent-mining
Key: deep_patent_mining
Official marketplace page: https://open.patsnap.com/marketplace/mcp-servers/patent-mining

Use evidence-backed technical topics, problems, effects, classifications, materials and applications to support—not replace—technical review.

Patent Monetization & Valuation — optional

Marketplace: https://open.patsnap.com/marketplace/mcp-servers/patent-monetize
Key: patent_monetization_valuation
Official marketplace page: https://open.patsnap.com/marketplace/mcp-servers/patent-monetize

Treat any result as one bounded input. Record method, assumptions, date, currency, data coverage and uncertainty. Never present a connector score as a transaction price.

No verified global PatSnap procurement connector is claimed. Procurement, company, finance, market, news, standards and legal evidence require separately authorized current sources.

Stage 1 — technology content analysis

1.1 Source and ownership register

Record:

provider and legal entity;
inventors/research team;
document provenance;
patent/application identifiers;
ownership and assignment evidence;
co-owners, funding obligations and university/employer rights;
licences, options, pledges, grants or field restrictions;
confidentiality/public-disclosure status; and
unresolved title or authority questions.

Do not equate named applicant with current ownership.

1.2 Technical summary

Provide:

one-sentence technical topic;
concise mechanism and method;
problem addressed;
conventional alternatives and limitations;
system boundary and dependencies;
four or more supported innovation points when evidence permits; and
three to five plausible application scenarios linked to requirements.

If the source is sparse, preserve unknowns rather than meeting a fixed word or item quota with invention.

1.3 Readiness assessment

Use TRL 1–9 only when appropriate to the discipline and supported by evidence. Record:

readiness framework and version;
claimed and independently supported level;
validation environment;
scale and sample count;
reproducibility;
safety/regulatory status;
manufacturing or implementation readiness;
supply-chain dependencies;
integration requirements; and
evidence gaps to reach the next level.

Do not convert “prototype,” “pilot,” or “field trial” mechanically into a TRL.

1.4 KPI register

For each KPI capture:

Field	Requirement
Name	Unambiguous metric
Value/range	Exact reported value
Unit	SI or stated domain unit
Method	Test protocol and conditions
Sample	Size, batch and comparator
Source	Evidence ID and locator
Uncertainty	Error, variability or limitation

Do not create a four-card KPI dashboard when fewer than four verified metrics exist.

1.5 Alternative comparison

Compare the technology with relevant alternatives across equivalent conditions. Include at least:

mechanism/architecture;
performance;
cost evidence or status;
scale/readiness;
integration/compatibility;
safety/regulatory burden;
supply-chain constraints; and
evidence date/source.

Use not comparable where test conditions differ materially. “Domestic substitution” is not a global comparison dimension; use supply security, regional availability and localization needs as relevant.

1.6 Transfer advantages and barriers

List evidence-backed:

recipient benefits;
integration fit;
time-to-value;
protected know-how or data;
training/support package;
validation assets;
scale-up barriers;
tacit-knowledge dependence;
required inventor participation; and
missing transfer package components.

Stage 2 — advancement and readiness assessment

Retain the source’s eight-factor baseline only as a configurable rubric:

Factor	Baseline weight
Breakthrough in technical principle	15
Novelty of core method	10
Difficulty of substitution	10
Height of technical barriers	10
Process/system compatibility	15
Commercialization readiness	15
Technical stability/reproducibility	10
Scalability potential	15

Before scoring:

confirm that the rubric fits the technology and transfer objective;
adjust weights if needed and record approval;
define anchors for 0, 25, 50, 75 and 100;
map evidence IDs to each factor;
identify missing evidence; and
record reviewer uncertainty.

For each factor output:

raw score;
weight;
weighted contribution;
evidence IDs;
one- or two-sentence rationale;
confidence;
missing evidence; and
sensitivity to a reasonable score range.

Suggested descriptive bands may be used only after disclosure:

0–39: early or insufficient evidence;
40–54: emerging;
55–69: differentiated;
70–84: strong evidence of advancement; and
85–100: exceptional under the declared rubric.

Do not call a technology “breakthrough” solely from the numerical band.

Stage 3 — patent and commercialization value review

3.1 Portfolio definition

Record:

search/entity scope;
patent population count and counting unit;
displayed subset and selection method;
simple/extended family rule;
applications versus grants;
jurisdictions;
status cutoff; and
ownership uncertainty.

3.2 Technical value

Assess with evidence:

claim/disclosure coverage of the transferable implementation;
dependence on unprotected know-how;
problem/means/effect relevance;
blocking or complementary relationships;
platform versus narrow-use potential;
citation context, not count alone; and
alternatives and design-around exposure.

3.3 Market value evidence

Assess:

addressable use cases;
recipient economics;
market-size source and methodology;
adoption barriers;
regulatory timing;
competitive substitutes;
geographic family alignment; and
credible commercialization evidence.

Never derive TAM from patent data alone.

3.4 Legal value evidence

Assess only with current records and qualified review where needed:

status by family member and as-of date;
remaining term assumptions;
claim version and scope;
prosecution/opposition/invalidation history;
ownership and encumbrances;
maintenance/annuity status;
territorial fit; and
known disputes.

Do not state “stable claims,” “valid patent,” or “no infringement risk” from database status alone.

3.5 Strategic value

Assess:

fit with recipient portfolios and roadmaps;
complementarity or gap filling;
standards relevance supported by standards/claim mapping;
licensing/assignment/joint-development feasibility;
bundling options;
know-how and data transfer; and
field/geography/exclusivity choices.

Do not label a patent an SEP candidate merely because its topic relates to a standard.

3.6 Valuation boundary

Remove the source’s fixed RMB ranges and scarcity/family/SEP multipliers.

When valuation is authorized, state:

valuation purpose;
method: cost, market, income, relief-from-royalty, option/scenario, or triangulation;
currency and valuation date;
revenue/cost/royalty assumptions;
probability and discount rates;
tax, term and territory;
comparable-transaction limitations;
low/base/high scenarios; and
sensitivity.

Otherwise output valuation_status: not_assessed and list required inputs.

3.7 Patent portfolio report modules

Preserve:

portfolio statistics with counting rules;
classification distribution with definitions;
selected-patent analysis with safe global links;
title, identifiers and dates;
abstract/claim-bounded interpretation;
classifications, status/as-of and evidence tags;
citation count with database/date/context;
filing chronology;
PCT/foreign filing opportunities, subject to deadlines and counsel; and
portfolio gaps and packaging recommendations.

Never force four IPC classes or a fixed patent count when evidence does not support them.

Stage 4 — potential-recipient discovery

4.1 Define recipient types

Candidate types may include:

operating companies with technical complementarity;
suppliers or integrators;
licensees in adjacent fields;
joint-development partners;
manufacturers or scale-up partners;
distributors with technical capability;
strategic acquirers;
spinout investors, when requested; and
public or nonprofit implementers.

Separate a company’s technical fit from willingness, authority, financial capacity and legal eligibility.

4.2 Candidate generation

Generate candidates from disclosed, auditable evidence such as:

relevant patent activity;
verified technical roadmaps;
product and capability evidence;
current investment, hiring or facility activity;
collaboration or licensing statements;
procurement opportunities;
grants and funded projects;
standards participation; and
supplied relationship data.

Do not infer purchase intent from one news article or one patent.

4.3 Entity resolution and eligibility

For every candidate confirm:

canonical legal name;
parent/subsidiary relationship;
operating region;
relevant business unit;
sanctions/export/control constraints where applicable;
conflicts and excluded parties;
current operational status; and
evidence IDs.

4.4 Three evidence dimensions

Retain the source baseline as configurable:

patent/technical evidence: 40 points;
current public/company evidence: 30 points; and
procurement/contract evidence: 30 points.

Procurement may be irrelevant in some sectors. In that case, redesign the rubric before scoring; do not silently redistribute missing points.

Use scripts/match_scorer.py for deterministic computation after evidence coding.

4.5 Patent/technical dimension

Candidate indicators include:

classification/topic overlap;
claim/disclosure semantic relevance;
recent relevant activity;
citation or family relationships interpreted in context;
complementary capabilities;
portfolio gap evidence; and
development-stage alignment.

Fixed thresholds such as three IPCs, 0.75 semantic similarity, 20% growth, 50 citations or five countries are not universal. Define sector-, age-, database- and geography-adjusted anchors.

4.6 Public/company signal dimension

Candidate indicators include:

explicit partnership or technology-acquisition intent;
facility/product expansion;
R&D program or laboratory investment;
university/government/consortium activity;
relevant hiring or capital allocation;
conference/standards participation; and
certification or regulatory progress.

Record event date, publication date, source quality and contradictory signals.

4.7 Procurement/contract dimension

Use only authorized, jurisdiction-appropriate sources. Candidate indicators include:

relevant open or recent procurement;
buyer/supplier role;
technical requirement fit;
budget and currency, normalized to the declared valuation date;
award history;
repeat/renewal behavior;
procurement stage and deadline; and
eligibility constraints.

The bundled scripts/gov_bid_api.py is a localized normalization adapter, not a configured global procurement API client. It must not send network requests or contain credentials.

4.8 Evidence coverage and missing data

For every metric record:

normalized value;
evidence IDs;
source quality;
as-of date;
confidence;
missing status; and
normalization anchor/version.

Missing is not zero. The scorer returns:

raw weighted score;
available-weight score;
evidence coverage;
confidence/quality indicator;
eligibility gate state; and
sensitivity interval.

Do not rank candidates below the approved evidence-coverage floor as if comparable.

4.9 Recipient ranking and cards

Each recipient card includes:

canonical company/entity name;
recipient type and relevant business unit;
total score and declared model version;
dimension scores and coverage;
evidence-quality label;
eligibility status;
at least three supported reasons when available;
contradictory or missing evidence;
one or two appropriate next validation actions; and
source IDs.

Use text grades, not star icons. Treat score bands as prioritization labels, not objective probabilities.

4.10 Robustness checks

Before final ranking:

vary top-level weights;
vary key normalization anchors;
test missing-data treatment;
compare rank stability;
inspect outliers;
check entity duplicates;
review source concentration; and
conduct a human reasonableness review.

Show unstable rankings as tiers or ranges.

Stage 5 — transfer risk and pathway

5.1 Four risk dimensions

Retain four 25-point baseline dimensions only after defining evidence anchors:

technical risk: readiness, reproducibility, key-person dependence, integration, scale-up and obsolescence;
market risk: demand, substitutes, adoption timing, pricing and policy/regulatory change;
legal/IP risk: title, scope, status, third-party rights, confidentiality, export/control and contract terms; and
partner risk: absorption capacity, financial/operational health, intent quality, governance and information asymmetry.

Risk score direction must be explicit: higher means greater risk.

5.2 Risk bands

If the user accepts the source baseline, use:

0–20: low;
21–40: moderate-low;
41–60: moderate-high;
61–80: high; and
81–100: critical.

These are rubric bands, not probabilities. Show each dimension’s evidence, uncertainty, owner, mitigation and decision gate.

5.3 Recipient-specific risk profiles

For the top three eligible recipients, provide unique profiles with:

integration mismatch;
licensing/ownership concern;
organizational or financial constraint;
geography/regulatory issue;
evidence gap;
mitigation action; and
validation owner.

Do not duplicate generic concerns across all cards.

5.4 Transfer pathway

Build a gated timeline rather than a guaranteed schedule:

near term, typically 0–6 months: evidence closure, ownership, package, outreach preparation, NDA and validation plan;
medium term, typically 6–18 months: evaluation, pilot, diligence, negotiation and milestones; and
longer term, typically beyond 18 months: scale-up, regulatory/quality work, commercialization, royalties/equity and post-transfer support.

Adapt timing to the sector.

Compare at least these routes when relevant:

licence;
assignment;
joint development;
sponsored research;
spinout/equity contribution; and
open/nonexclusive access.

For each route show control, exclusivity, funding, IP improvements, data/know-how, territory/field, milestones, termination, audit and support implications.

Stage 6 — report generation and file service

Use scripts/report_generator.py to create the static report only after hard validation passes.

The report must contain:

executive decision summary;
scope, methods, evidence cutoff and data coverage;
technology content and transfer package;
readiness and KPI evidence;
alternative comparison;
advancement assessment;
patent/commercialization value review;
portfolio statistics and selected patents;
recipient-discovery method and ranking;
recipient cards and sensitivity;
four-dimension risk review;
individualized top-recipient risk profiles;
pathway and route comparison;
source register; and
limitations and next validation actions.

The renderer must:

escape all untrusted values;
allow only safe HTTP(S) links;
be self-contained and work without a CDN;
use semantic English HTML;
use accessible native SVG/tables instead of opaque external charts;
expose data values in text;
use color-independent labels;
support keyboard, mobile and print; and
never open a browser or server automatically.

Use scripts/report_generator_patch.py only to write to an exact user-approved path and optional user-approved backup path.

Do not:

write to Desktop, home, session or temp by assumption;
overwrite an existing file without explicit permission;
start an HTTP server;
announce a path that was not successfully written; or
create files inside this skill package at runtime.

Validation gate

Stop report generation if any hard check fails:

technology identity/scope unresolved;
provider or ownership assertion unsupported;
confidential handling not approved;
invented KPI, comparison, score, patent, market, company or procurement fact;
recipient entity unresolved;
prohibited recipient or compliance gate failed;
score lacks evidence IDs;
evidence coverage below the approved ranking floor;
missing data treated as zero or silently renormalized;
currency/date/unit absent from money values;
unsafe URL or unescaped text;
valuation presented without method and assumptions;
legal/SEP/FTO conclusion presented without proper evidence/review; or
output path not approved.

Failure paths

Patent connectors unavailable

Deliver a non-executed portfolio/search plan and required query/entity fields. Do not score patent evidence.

Procurement source unavailable or irrelevant

Mark the dimension not_scored. Redesign/approve an alternative rubric if recipient ranking must continue.

Sparse technology evidence

Deliver a transfer-package gap analysis and evidence-acquisition plan before recipient ranking.

Ambiguous ownership

Pause commercialization conclusions and route title/authority review to qualified counsel or the technology-transfer office.

Insufficient candidate evidence

Return an unranked candidate longlist with evidence gaps. Do not manufacture three reasons or an action.

Valuation inputs missing

Set not_assessed. Provide a data request and scenario template.

Unstable ranking

Use tiers/ranges, show weight sensitivity and identify decisive missing evidence.

Renderer or output failure

Do not claim delivery. Return the exact error, preserve validated input, and retry only within the approved path.

Quality checklist

Both source input modes remain supported with a safe Mode B gate.
All six source stages are present.
All eight advancement factors are preserved as a configurable baseline.
All four patent/commercialization value dimensions are present.
All three recipient evidence dimensions and two-layer logic are represented.
All four transfer-risk dimensions and top-three individualized profiles are present.
The three time horizons and transfer routes are present.
Every metric, score, recommendation and risk maps to evidence.
Missing data and coverage are explicit.
Patent, company, market, procurement and valuation semantics are separated.
MCP names and links are verified global services only.
No source credential, China-only provider assumption or regional domain remains.
Report design is English, scientific, accessible, static and print-safe.
Output writes occur only to user-approved paths.