Fully autonomous AI Agents system capable of performing complex penetration testing tasks
Go ai-agentsai-security-toolanthropicautonomous-agentsgolanggptgraphqlmulti-agent-systemoffensive-securityopen-sourceopenaipenetration-testingpenetration-testing-toolsreactsecurity-automationsecurity-testingsecurity-toolsself-hosted
test(editor): corpus asserts a word multiset, not a Set
Sergey Kozyrenko1386f5e747df
diff --git a/frontend/src/components/shared/markdown-editor-corpus.test.ts b/frontend/src/components/shared/markdown-editor-corpus.test.tsindex d8b76825..08c184c0 100644--- a/frontend/src/components/shared/markdown-editor-corpus.test.ts+++ b/frontend/src/components/shared/markdown-editor-corpus.test.ts@@ -39,9 +39,9 @@ describe('corpus — every real prompt .tmpl survives the round-trip with no conexpect(save1).not.toContain('<');expect(save1).not.toContain('>');expect(variables(save1)).toEqual(variables(src));- const after = new Set(words(save2));- const lost = [...new Set(words(src))].filter((w) => !after.has(w));- expect(lost).toEqual([]);+ // word MULTISET (not a Set) so a count drop / dup / reorder is caught, plus ≤2-save convergence+ // (some templates canonicalize over two saves — save1 ≠ save2 is within contract, save2 is stable)+ expect(words(save2).slice().sort()).toEqual(words(src).slice().sort());expect(roundTrip(save2)).toBe(save2);});}