The authors represent a repository as a dependency graph, collapse dependency cycles into a directed acyclic graph, and compute content hashes for each module's dependency cone.
Code-linked evidence lets AI software claims expire when dependencies change
Assay binds test, security and review claims to dependency-aware hashes, then mechanically rejects stale or insufficient evidence at merge time.
Big Tech
Om Shankar Tiwari · Tangi Vass · Gagan Deep Singh
Google · OMNI3ai · Glinr Studios · theSVG.org
Research Digest··3 min read
Tiwari, Vass and Singh present Assay, a Python system that combines repository indexing with an evidence ledger for AI-assisted software delivery.
Why this paper
From Google and 3 others · Released code
In one line
Assay binds agent claims to Merkle hashes of dependency cones so claims expire automatically when code changes.
What it released
Code
What we could check
- ✓Code link in the paper (github.com)
- ·No weights link found
- ·No dataset link found
- ·No compute details found
- ✓Limitations stated by the authors
- ✓Reports numbers on named benchmarks (4 benchmarks)
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§