Competitive Differentiation Assessment

Purpose

Interpret competitive landscape and patent search results for a concept, producing structured findings that separate three independent dimensions. This skill applies after oase patent search has been run and patent data is available.

When to use

• Before committing to Cohort B characterization (per Rule 16 in the lead template: "Screen for competitive landscape and FTO before structural work.")
• When assessing competitive positioning of a tournament-recommended idea.
• When Hypex generation/reflection identifies a target that may have existing competitor activity.

The three dimensions

These three dimensions must never be collapsed into a single score. Each is reported separately, with its own search metadata:

1. Competitor activity

What programs, compounds, publications, and patent filings exist for this target, modality, or indication combination.

• Not a veto. A crowded field with a genuinely differentiated angle should surface as "differentiated despite crowding", not be rejected by a naive "many competitors = bad" heuristic.
• Record: density (uncrowded / low_activity / moderate_activity / high_activity), top assignees, recent patent count.

2. Patentability / novelty

Is there white space for new intellectual property?

• Assessed by overlap. How many existing patents cover the same modality + indication combination?
• Record: novelty_assessment (high_novelty / potential_novelty / partial_overlap / crowded_space).

3. Freedom to operate (FTO)

Can the concept be pursued without infringing existing intellectual property?

• Never implies legal clearance. Every FTO finding carries the mandatory disclaimer (see below).
• Record: risk_level (no_recent_filings / low_risk / moderate_risk / high_risk), jurisdictions found, unresolved questions.

Mandatory disclaimers

FTO disclaimer (NEVER omit)

This assessment is based on public patent database searches and publicly available information. It is NOT formal legal clearance. A public search or structural similarity analysis cannot substitute for a formal freedom-to-operate opinion by qualified patent counsel. Material FTO conclusions require qualified legal review.

This disclaimer must appear on every FTO-related finding, assessment record, and analysis output. It is not optional prose.

Coverage limits (NEVER omit)

Every search finding must state:

1. What was searched (search scope) — e.g., "Google Patents public search for 'CDK4 inhibitor'".
2. What was NOT searched (coverage limits) — e.g., "Non-English filings may be underrepresented; unpublished applications within 18-month window are not visible; patent databases other than Google Patents were not queried."
3. The search date — results reflect the database state at this date.

No patent search finding may be presented as exhaustive.

Mapping to assessment records

The differentiation assessment produces oase.evidence-assessment.v1 records with the following evidence types:

Evidence type naming convention

Following the design doc (SS1.5): lowercase, underscored, descriptive of the evidence domain rather than the specific tool.

• competitive_precedent — ligand, program, or patent precedent in the competitive landscape.
• patent_landscape — patent filing landscape relevant to IP position.

Workflow

1. Run `oase patent search "<target>"` to fetch patent data.
2. Run `oase differentiation assess <target> --concept IC-NNN
       --modality <modality> --indication <indication>`.
3. Review the three-dimension output.
4. Run `oase differentiation report <target> --concept IC-NNN`
   to generate assessment records.
5. Write the assessment records to the control store using
   `controlstore.write_record()`.

Integration with Hypex

When Hypex generation or reflection produces hypotheses about a target:

1. The generation agent should use oase patent search as part of its Phase 1 broad exploration to check for existing IP.
2. The reflection agent should check patent landscape as part of its Phase 2c prior art search.
3. The competitive differentiation assessment complements Hypex's scientific evaluation — Hypex evaluates scientific merit, this skill evaluates competitive positioning.

The two are complementary and should not duplicate each other.

Charter constraints

When the charter specifies program-level constraints (e.g., "no oral formulations" or "oncology indications only"), these are recorded as separate assessment records with evidence_type: competitive_precedent and a rationale that explicitly identifies them as program constraints rather than scientific rejections.

A charter constraint is a program-specific exclusion, distinct from:

• A scientific rejection (evidence contradicts the hypothesis)
• A competitive rejection (too many competitors)
• An FTO concern (blocking IP exists)

Reconciliation with Cohort A FTO wording

The lead template (science-program-lead/agents.md, section 9) describes Cohort A step 3: "Competitive landscape / FTO screen." The new tooling's disclaimers must match the lead template's standard:

• The lead template's "If step 3 reveals blocking IP with no white space: terminate" is a decision by the lead, not an automated gate.
• The differentiation tool surfaces information; it does not make the terminate/proceed decision.
• Both upstream (lead template) and downstream (this skill) agree: a public search is not formal legal clearance.