research-agenda

Research Agenda

Create a reviewed agenda of work worth doing. Do not treat every gap as a contribution.

Read First

• references/repository-contract.md
• references/workflow-stage-contracts.md
• references/research-positioning-policy.md

Workflow

1. Run npm run workflow:agenda when available.
2. Read research_agenda/agenda-contract.md, sota/gaps.md, sota/sota-claim-ledger.csv, and survey/survey-claim-ledger.csv.
3. Add candidate opportunities to research_agenda/opportunity-ledger.csv.
4. Evaluate novelty, feasibility, evidence, expected contribution, failure condition, risk, cost, publishability, and release constraints.
5. Write direction records in research_agenda/directions/ only for opportunities that survive initial review.
6. Store final agenda synthesis in research_agenda/final/.

Review Loop

Review each opportunity adversarially, fix weak framing or evidence gaps, and re-review until the decision is accepted, rejected, or held with a clear rationale.

Handoff

Hand off accepted opportunities to contribution-package with linked SOTA and survey claim IDs.