Research Roadmap

Maintained by the Strategist agent. Updated: 2026-07-24

Active Investigations

1. Hop Calculation Algorithms

Why: The graduated trust model depends on computing shortest meaningful paths through a signed attestation graph. Known algorithms (EigenTrust, PageRank variants) need adaptation for revocable, scoped, expiring edges.

Questions:

How does EigenTrusts transitive trust calculation perform with 10,000 nodes?

What does shortest-path-with-path-quality-weighting look like algorithmically?

How does the graph recompute when edges expire or are revoked?

Source: Graduated Trust - Hop calculation subtleties

2. Sybil Resistance at Graph Edges

Why: Open question in the graduated trust proposal. A cluster of mutually-vouching fake accounts could gain trust proximity through strategic positioning.

Questions:

What prevents a Sybil attack on a web-of-trust graph?

Does requiring disjoint paths (at least N independent vouchers) provide sufficient resistance?

Are there known bounds from academic literature on Sybil-resistant trust graphs?

Source: Graduated Trust - Open Questions

3. Privacy-Preserving Graph Traversal

Why: A fully public trust graph reveals contributor relationships that may be sensitive. Some contributors may not want their relationship structure public.

Questions:

Can hop distance be computed without revealing the full graph to the querier?

Are zero-knowledge approaches practical for this use case?

Does selective disclosure (revealing path only to the project) provide adequate privacy?

Source: Graduated Trust - Open Questions

4. Reputation Calibration and Recovery from Compromise

Why: How severe should reputation degradation be when a vouch goes bad? How does the graph recover when a high-reputation node is compromised?

Questions:

What is the right decay function for reputation loss by hop distance?

Can recovery be bootstrapped from out-of-band verification?

Is there a time-based healing mechanism that prevents permanent reputation loss?

Source: Graduated Trust - Open Questions and Recovery

Future Investigations

5. Cross-Domain Governance Models

**Who decides what cryptography scope means? The domain governance question is open and likely needs community-specific answers rather than protocol-level enforcement.

6. Proof-of-Context Question Generation

**Can LLMs generate meaningful context questions from git history? Can they evaluate answers reliably? This is the most experimentally testable mechanism in the proposal.

7. Existing Trust Tooling Landscape

**What existing tools approach this problem? Sigstore, in-toto, TUF, SPIFFE/SPIRE. How does graduated trust differ, and where does it complement?

Research Principles

Prefer mechanisms that are implementable over ones that are provably perfect

Document open questions clearly so others can contribute

Prioritize investigations that unblock the top-ranked opportunities in the Opportunity Backlog

Every research finding should trace back to a specific question in a published document