Changelog

v0.29

Current Anvil and Bifrost runtime

• Bundled scans now run Anvil 0.21.2 and Bifrost 0.8.0, keeping ACP execution and analyzer tooling aligned with their latest compatible published releases.
• The server image now builds with Rust 1.96 to support the upgraded analyzer runtime.

v0.28.2

Large repository memory guardrail

• The default degraded subset cap is now 10,000 files, reducing peak memory pressure on extremely large repositories.

v0.28.1

Large repository migration hotfix

• The production database migration for large-repository scan scope now applies as a forward-only schema change without modifying an already-applied baseline migration.

v0.28

Large repository scan safeguards

• Oversized repositories now continue in a degraded subset mode instead of exhausting analyzer memory after checkout.
• Scans that exceed the file limit clearly show that only a subset was analyzed and are kept out of public scoring.
• The default analysis cap is now 30,000 files, which protects production workers from repeatedly restarting on extremely large repositories.

v0.27

ACP specialist reliability, safer publishing, and clearer reports

• ACP scans now run specialist work with stronger diagnostics, model-fallback handling, and recovery paths for empty provider output.
• Specialist orchestration now uses fuller lane rosters, clearer Bifrost tool names, and improved static-analysis feed behavior after terminal scan states.
• Bundled Docker builds now use the latest tagged Anvil and Bifrost releases for ACP execution and analyzer tooling.
• Closed case files now render lane summaries, scorecard notes, inline markdown, file pills, and code references more cleanly across narrow layouts.
• Previous-report comparison and deduplication now produce steadier case-history context for repeat scans.
• Most Wanted publishing now has an additional safety gate so reports are checked before public leaderboard exposure.
• The backend schema setup now uses a SQLx baseline, making production migrations and future schema changes easier to validate.

v0.26

Rust ACP analysis, model tooling, and cleaner reports

• Case files now include clearer previous-report comparisons so repeat scans can show what changed since the last report.
• Recovered specialist evidence and partial scorecards now stay better aligned with the visible case file, reducing confusing recovered-lane presentation.
• Repository identity is preserved more consistently across public scans, Most Wanted entries, and shared case links.
• Report synthesis now keeps claims more tightly scoped to accepted evidence, with safer handling when a category cannot be scored.
• Scan execution now uses the newer ACP-backed analysis path, improving specialist orchestration and making model configuration easier to validate.
• Score review and sniff-test stages were tightened so valid reports are less likely to be degraded by downstream structured-output handling.
• Specialist cards now avoid duplicate headline rendering, making closed case files easier to scan.

v0.25

Printable reports, safer scoring, and clearer case history

• Case files and Most Wanted entries now have dedicated print routes with printable report actions, component-aware report rendering, and cleaner case identifiers.
• Reports now handle incomplete specialist evidence more clearly, with better scoped non-finding language and safer partial scoring when some categories cannot be assessed.
• Most Wanted now groups repository history more clearly, supports previous-scan comparison cues, and adds feature-request intake.

v0.24

More complete reports and stronger specialist evidence

• Specialist findings are now captured in a stricter structured format, making completed reports more consistent and easier to compare.
• Final reports are less likely to silently lose important evidence when one specialist lane is incomplete.
• Failed or incomplete specialist lanes are isolated more clearly so useful findings from other lanes can still appear in the case file.

v0.23

Lane recovery, evidence discipline, and safer partial reports

• Partial specialist results are now handled more consistently, reducing cases where one weak lane prevents the rest of the report from being useful.
• Reports now distinguish confirmed findings, tentative leads, and scoped non-findings more carefully.
• Incomplete evidence is shaped more safely in public reports, with clearer limits when a lane could not support a full conclusion.

v0.22

Provider expansion, previous-report review, and safer scan outcomes

• Public scans now support provider-aware GitLab and Bitbucket repositories alongside GitHub, including improved repository identity handling, public repository metadata, and Most Wanted social previews.
• New scans can review previous reports before generating fresh findings, giving synthesis and lane processing more context when repositories have already been inspected.
• Partial and no-score scan outcomes are now handled more explicitly, with clearer public messages when a report cannot be fully scored.
• Scan intake and report flows now include stronger privacy and retention hints, repo-specific metadata in closed reports, and new footer and Discord support affordances.
• Infrastructure-heavy repositories are now classified more reliably so reports stay better matched to the kind of project being scanned.