@memnexus-ai/mx-agent-cli 0.1.242 → 0.1.243

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
@@ -0,0 +1,21 @@
1
+ /**
2
+ * Tests for #4707 — the behavioral block is read from origin/<baseBranch>, the
3
+ * delivered CLAUDE.local.md is verified for CONTENT, and repeated markers fail
4
+ * loudly.
5
+ *
6
+ * WHAT WENT WRONG. `mx-agent start` resolved the team catalog and the roleguide
7
+ * from `projectRoot` — `/workspace`, the main repo working tree — which was 82
8
+ * commits behind origin/main. The markers eval's roleguide gained in #4521 were
9
+ * in git and had never been read, so eval shipped an isolation-only
10
+ * CLAUDE.local.md for four weeks while the same startup sequence printed
11
+ * "Agent config generated from … (origin/main)".
12
+ *
13
+ * HOW THESE TESTS ARE PINNED. Every git-vs-local test builds a repository where
14
+ * the two sources DISAGREE: origin/<branch> carries one roleguide, the
15
+ * projectRoot working tree carries a different one. Asserting on the CONTENT
16
+ * (not merely on "a block exists") means reverting the read to projectRoot fails
17
+ * the test rather than passing it by luck — which is exactly how the previous
18
+ * iteration's verification passed while the outcome it claimed was false.
19
+ */
20
+ export {};
21
+ //# sourceMappingURL=behavioral-git-read.test.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"behavioral-git-read.test.d.ts","sourceRoot":"","sources":["../../src/__tests__/behavioral-git-read.test.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;GAkBG"}