weichen/Agent-System-Backend
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
473 Commits 17 Branches 11 Tags
c50f9f07e833dc18264d547cfbfd67eea565ac62
Commit Graph
5 Commits
Author SHA1 Message Date
weichen c50f9f07e8 Guard source previews against content version substitution 
Why: Opening a processed source must not show changed current bytes as the quoted version.

Evidence: SELF TaskSourceContent regression fails before and passes after expected-hash enforcement; artifact regressions and live compilation pass.

Impact: Optional content hash narrows the existing authorized artifact read; controlled fixtures select a genuinely provided quotation.

Refs: fast-slow-u2-u3 source-version acceptance
 2026-10-13 02:51:07 +00:00
weichen c2851d498f Use repository prompt recipes in controlled model comparison 
Why: Baseline and candidate must exercise their actual default prompts with identical route and resource limits.

Evidence: SELF live-tag compilation passes.

Impact: Evaluation harness only; normal automatic notice quality remains outside the controlled comparison.

Refs: fast-slow-u2-u3 Task 7
 2026-10-13 02:41:16 +00:00
weichen 4184430d6d Keep source changes visible and bound evaluation calls consistently 
Why: Updated or missing sources must remain visible as limitations, and comparison calls need stable prompt grouping.

Evidence: SELF card/server suites and live-tag compilation pass.

Impact: Adds neutral source limitations and refines test-only provider accounting.

Refs: fast-slow-u2-u3 Tasks 6 and 7
 2026-10-13 02:36:57 +00:00
weichen d7597fea05 Build the isolated evaluator with the native provider client 
Why: The live harness must use the actual client constructor and preserve MiniMax reasoning separation.

Evidence: SELF live-tag compilation passes without provider calls.

Impact: Test harness only; no production route changes.

Refs: fast-slow-u2-u3 Task 7
 2026-10-13 02:31:10 +00:00
weichen 13e880feec Publish validated material findings during background work 
Why: Real source-backed discoveries should reach the same task card before the slow channel completes.

Evidence: SELF receipts record runtime and manager checkpoint red-green tests, scoped source/version checks, privacy filtering, quota corrections, and SQLite/PG regressions; fixed composition and live comparison follow.

Impact: Adds a local material-only checkpoint sink with no summary model calls, uses version witnesses for source commits, and preserves existing reading and notification paths.

Refs: fast-slow-u2-u3 Tasks 6 and 7; test-only live HTTP harness bounds shared provider requests without approving HITL
 2026-10-13 02:29:23 +00:00