@teevee98/synapse-hub 4.3.1 → 4.4.1
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.
- package/.cursor/rules/agents.mdc +11 -11
- package/.cursor/rules/katalon.mdc +11 -0
- package/.cursor/rules/perf.mdc +10 -0
- package/.cursor/rules/playwright.mdc +10 -0
- package/.cursor/rules/pytest.mdc +10 -0
- package/.cursor/rules/synapse-qa.mdc +4 -2
- package/.cursorrules +3 -10
- package/.windsurfrules +3 -10
- package/AGENTS.md +4 -3
- package/CLAUDE.md +4 -5
- package/README.en.md +3 -3
- package/README.md +3 -3
- package/agentskills.json +15 -15
- package/{.agents/agents → ai-agents/personas}/api-engineer.md +9 -1
- package/{.agents/agents → ai-agents/personas}/failure-analyst.md +1 -1
- package/{.agents/agents → ai-agents/personas}/live-explorer.md +1 -1
- package/{.agents/agents → ai-agents/personas}/orchestrator.md +8 -15
- package/{.agents/agents → ai-agents/personas}/perf-engineer.md +11 -5
- package/{.agents/agents → ai-agents/personas}/pom-crafter.md +3 -9
- package/{.agents/agents → ai-agents/personas}/quality-gatekeeper.md +1 -1
- package/{.agents/agents → ai-agents/personas}/test-designer.md +2 -1
- package/{.agents/agents → ai-agents/personas}/test-runner.md +3 -4
- package/{.agents/rules → automation/playwright}/mandatory-playwright-live.md +5 -13
- package/{.agents/skills/01-web-automation → automation/playwright}/playwright-webapp-testing/SKILL.md +2 -0
- package/{.agents/skills/01-web-automation → automation/shared-utils}/clean-pom-crafter/SKILL.md +14 -3
- package/{.agents/skills/01-web-automation → automation/shared-utils}/e2e-testing-patterns/SKILL.md +12 -1
- package/{.agents/skills/01-web-automation → automation/shared-utils}/lam-script-generator/SKILL.md +2 -0
- package/docs/ARCHITECTURE.en.md +96 -3
- package/docs/ARCHITECTURE.md +87 -3
- package/docs/TOOL_CATALOG.md +17 -5
- package/docs/reports/AUDIT_REPORT.md +1038 -0
- package/docs/reports/PROJECT_HEALTH_REPORT.md +49 -0
- package/docs/reports/TECHNICAL_DEBT_LEDGER.md +22 -0
- package/package.json +8 -1
- package/plugin.json +1 -1
- package/{.agents/skills/02-test-design → requirement/spec-review}/review-spec/SKILL.md +8 -0
- package/scripts/install_global_plugin.ps1 +1 -1
- package/scripts/install_global_plugin.sh +1 -1
- package/skills_index.json +61 -0
- package/skills_manifest.json +88 -62
- package/{.agents/skills/08-code-quality → standards/coding-standards}/lean-code-ladder/SKILL.md +8 -1
- package/{.agents/skills/08-code-quality → standards/coding-standards}/surgical-coding-principles/SKILL.md +2 -0
- package/{.agents/skills/11-ai-and-prompting → standards/context-ai}/mipro-prompt-compiler/SKILL.md +8 -0
- package/synapse.config.json +1 -1
- package/templates/ai_bridges/CLAUDE.md +4 -5
- package/templates/ai_bridges/copilot-instructions.md +1 -0
- package/templates/ai_bridges/cursor_synapse_qa.mdc +4 -2
- package/templates/ai_bridges/windsurfrules +3 -10
- package/templates/test-plan/performance-test-plan-template.md +106 -0
- package/{.agents/skills/02-test-design → test-design/manual-rbt}/rbt-manual-testing/SKILL.md +11 -5
- package/{.agents/skills/07-data-and-state → test-design/test-data}/file-and-report-validator/SKILL.md +8 -1
- package/{.agents/skills/07-data-and-state → test-design/test-data}/synthetic-test-data-generator/SKILL.md +4 -1
- package/testing/api-contracts/api_testing_rules.md +74 -0
- package/{.agents/skills/06-healing-and-resilience → testing/healing-resilience}/agentless-qa-repair/SKILL.md +2 -0
- package/{.agents/skills/06-healing-and-resilience → testing/healing-resilience}/failure-autopsy-and-healing/SKILL.md +5 -11
- package/{.agents/skills/06-healing-and-resilience → testing/healing-resilience}/self-healing-locator/SKILL.md +2 -0
- package/testing/manual/manual-qa-and-rtm-bridge/SKILL.md +92 -0
- package/testing/manual/manual-qa-and-rtm-bridge/examples/sample_bug_and_rtm.json +38 -0
- package/{.agents/skills/04-performance-and-load → testing/performance}/backend-load-and-stress-patterns/SKILL.md +5 -2
- package/testing/performance/backend-load-and-stress-patterns/references/performance_test_plan_29119.md +106 -0
- package/{.agents/rules → testing/performance}/performance_testing_rules.md +27 -3
- package/{.agents/skills/04-performance-and-load → testing/performance}/web-vitals-and-performance-profiler/SKILL.md +75 -14
- package/{.agents/skills/05-security-and-compliance → testing/security-a11y}/security-dast-sentinel/SKILL.md +13 -2
- package/{.agents/rules → testing/security-a11y}/security_testing_rules.md +19 -0
- package/{.agents/skills/05-security-and-compliance → testing/security-a11y}/wcag-audit-patterns/SKILL.md +4 -1
- package/tools/.infer_baseline.json +44 -3
- package/tools/TOOLS_RUNBOOK.md +246 -4
- package/tools/TOOL_CATALOG.json +199 -19
- package/tools/TOOL_INDEX.json +160 -0
- package/tools/__init__.py +1 -1
- package/tools/ai_eval_sentinel.py +403 -0
- package/tools/assertion_optimizer.py +188 -0
- package/tools/asvs_auditor.py +410 -0
- package/tools/author_integrity.py +12 -1
- package/tools/benchmark_swarm_context.py +1 -1
- package/tools/browser_helpers.js +2 -2
- package/tools/catalogs/api_contract_testing.json +107 -0
- package/tools/catalogs/auxiliary_utilities.json +1536 -0
- package/tools/catalogs/benchmarking.json +23 -0
- package/tools/catalogs/ci_cd_packaging.json +46 -0
- package/tools/catalogs/code_generation_ast.json +70 -0
- package/tools/catalogs/failure_autopsy_healing.json +41 -0
- package/tools/catalogs/manual_testing_design.json +80 -0
- package/tools/catalogs/orchestration.json +25 -0
- package/tools/catalogs/performance_testing.json +71 -0
- package/tools/catalogs/quality_gates.json +65 -0
- package/tools/catalogs/sandboxing_isolation.json +42 -0
- package/tools/catalogs/security.json +72 -0
- package/tools/catalogs/setup_onboarding.json +26 -0
- package/tools/catalogs/testing_utilities.json +111 -0
- package/tools/chaos_resilience_probe.py +362 -0
- package/tools/check_env.py +16 -3
- package/tools/check_version_sync.py +114 -63
- package/tools/cldr_l10n_verifier.py +309 -0
- package/tools/cli.py +971 -3
- package/tools/core/circuit_breaker.py +106 -0
- package/tools/core/token_estimator.py +56 -7
- package/tools/cvss_calculator.py +412 -0
- package/tools/dashboard_generator.py +12 -12
- package/tools/etl_data_reconciler.py +460 -0
- package/tools/generate_jmx_plan.py +12 -4
- package/tools/generate_tool_catalog.py +60 -2
- package/tools/jtl_analyzer.py +23 -1
- package/tools/mcp_docx_server.py +1 -1
- package/tools/mcp_excel_server.py +1 -1
- package/tools/mcp_server.py +5 -1
- package/tools/mobile_resilience_sentinel.py +273 -0
- package/tools/observability_tester.py +235 -0
- package/tools/payment_interop_simulator.py +337 -0
- package/tools/pre_push_guard.py +306 -0
- package/tools/project_state.py +17 -10
- package/tools/regulatory_compliance.py +336 -0
- package/tools/reversible_sandbox.py +7 -2
- package/tools/rules_router.py +6 -4
- package/tools/runaway_stopper.py +19 -111
- package/tools/setup_project.py +1 -1
- package/tools/skill_canonical_digest.py +2 -2
- package/tools/skillspector.py +78 -10
- package/tools/sync_agent_rules.py +15 -27
- package/tools/sync_skills_manifest.py +38 -4
- package/tools/tmmi_auditor.py +179 -0
- package/tools/ui_server.py +1 -1
- package/tools/update_sentinel.py +4 -4
- package/tools/validate_host_interop.py +3 -1
- package/tools/validate_skills_schema.py +49 -2
- package/tools/vpat_exporter.py +428 -0
- package/tools/welcome_page.py +4 -4
- package/.agents/rules/stack-auto-router.md +0 -30
- package/.agents/skills/09-ci-and-devops/github-actions-templates/SKILL.md +0 -114
- package/.agents/skills/09-ci-and-devops/github-actions-templates/examples/matrix_test_pipeline.yml +0 -9
- package/.agents/skills/09-ci-and-devops/gitlab-ci-templates/SKILL.md +0 -123
- package/.agents/skills/09-ci-and-devops/gitlab-ci-templates/examples/sample_gitlab_ci.yml +0 -17
- package/.agents/skills/09-ci-and-devops/jenkins-pipeline-templates/SKILL.md +0 -133
- package/.agents/skills/09-ci-and-devops/jenkins-pipeline-templates/examples/Jenkinsfile.example +0 -21
- package/.cursor/rules/language.mdc +0 -7
- package/templates/ci/github-actions-synapse.yml +0 -95
- package/templates/ci/gitlab-ci-synapse.yml +0 -16
- package/tools/build_consolidated_report.py +0 -1643
- package/tools/create_mock_docx.py +0 -107
- /package/{.agents → ai-agents/hooks}/hooks.json +0 -0
- /package/{.agents → ai-agents}/hooks/session_start.py +0 -0
- /package/{.agents/skills/10-agent-orchestration → ai-agents/orchestration}/actor-qa-swarm/SKILL.md +0 -0
- /package/{.agents/skills/10-agent-orchestration → ai-agents/orchestration}/actor-qa-swarm/examples/broker_sharding_sample.json +0 -0
- /package/{.agents/skills/10-agent-orchestration → ai-agents/orchestration}/agent-orchestration-multi-agent-optimize/SKILL.md +0 -0
- /package/{.agents/skills/10-agent-orchestration → ai-agents/orchestration}/agent-orchestration-multi-agent-optimize/examples/swarm_workload_config.json +0 -0
- /package/{.agents/skills/10-agent-orchestration → ai-agents/orchestration}/event-stream-qa-orchestrator/SKILL.md +0 -0
- /package/{.agents/skills/10-agent-orchestration → ai-agents/orchestration}/event-stream-qa-orchestrator/examples/timeline_sample.jsonl +0 -0
- /package/{.agents/skills/10-agent-orchestration → ai-agents/orchestration}/persistent-planning-files/SKILL.md +0 -0
- /package/{.agents/skills/10-agent-orchestration → ai-agents/orchestration}/persistent-planning-files/examples/task_plan_example.md +0 -0
- /package/{.agents/skills/10-agent-orchestration → ai-agents/orchestration}/qa-memory-os/SKILL.md +0 -0
- /package/{.agents/skills/10-agent-orchestration → ai-agents/orchestration}/qa-memory-os/examples/archival_memory_sample.json +0 -0
- /package/{.agents → ai-agents}/workflows/graphify.md +0 -0
- /package/{.agents → ai-agents}/workflows/test-impact.md +0 -0
- /package/{.agents → ai-agents}/workflows/verify.md +0 -0
- /package/{.agents/rules → automation/cypress}/cypress_rules.md +0 -0
- /package/{.agents/skills/01-web-automation → automation/katalon}/katalon-cli-runner/SKILL.md +0 -0
- /package/{.agents/skills/01-web-automation → automation/katalon}/katalon-cli-runner/examples/headless_kre_exec.sh +0 -0
- /package/{.agents/skills/01-web-automation → automation/katalon}/katalon-cli-runner/references/groovy-patterns.md +0 -0
- /package/{.agents/skills/01-web-automation → automation/katalon}/katalon-cli-runner/references/katalon-api.md +0 -0
- /package/{.agents/skills/01-web-automation → automation/katalon}/katalon-cli-runner/references/runner_commands.md +0 -0
- /package/{.agents/skills/01-web-automation → automation/mobile}/mobile-appium-som/SKILL.md +0 -0
- /package/{.agents/skills/01-web-automation → automation/mobile}/mobile-appium-som/examples/BaseScreen.java +0 -0
- /package/{.agents/skills/01-web-automation → automation/mobile}/mobile-appium-som/examples/LoginScreen.java +0 -0
- /package/{.agents/skills/01-web-automation → automation/mobile}/mobile-appium-som/examples/TC_MobileLogin.java +0 -0
- /package/{.agents/skills/01-web-automation → automation/mobile}/mobile-maestro-detox-patterns/SKILL.md +0 -0
- /package/{.agents/skills/01-web-automation → automation/mobile}/mobile-maestro-detox-patterns/examples/sample_flow.yaml +0 -0
- /package/{.agents/rules → automation/mobile}/mobile_appium_rules.md +0 -0
- /package/{.agents/skills/01-web-automation → automation/playwright}/playwright-live-interaction/SKILL.md +0 -0
- /package/{.agents/skills/01-web-automation → automation/playwright}/playwright-live-interaction/examples/live_probe_script.js +0 -0
- /package/{.agents/skills/01-web-automation → automation/playwright}/playwright-live-interaction/references/advanced_superpowers.md +0 -0
- /package/{.agents/skills/01-web-automation → automation/playwright}/playwright-live-interaction/references/compact_accessibility_snapshots.md +0 -0
- /package/{.agents/skills/01-web-automation → automation/playwright}/playwright-live-interaction/references/recipes.md +0 -0
- /package/{.agents/skills/01-web-automation → automation/playwright}/playwright-live-interaction/references/swarm_orchestration.md +0 -0
- /package/{.agents/skills/01-web-automation → automation/playwright}/playwright-webapp-testing/examples/server_lifecycle_playwright.py +0 -0
- /package/{.agents/skills/01-web-automation → automation/shared-utils}/auth-session-and-cookie-manager/SKILL.md +0 -0
- /package/{.agents/skills/01-web-automation → automation/shared-utils}/auth-session-and-cookie-manager/examples/multi_role_auth_matrix.md +0 -0
- /package/{.agents/skills/01-web-automation → automation/shared-utils}/auth-session-and-cookie-manager/examples/storage_state_sample.json +0 -0
- /package/{.agents/skills/01-web-automation → automation/shared-utils}/bats-testing-patterns/SKILL.md +0 -0
- /package/{.agents/skills/01-web-automation → automation/shared-utils}/bats-testing-patterns/examples/cli_runner_test.bats +0 -0
- /package/{.agents/skills/01-web-automation → automation/shared-utils}/bats-testing-patterns/resources/implementation-playbook.md +0 -0
- /package/{.agents/skills/01-web-automation → automation/shared-utils}/clean-pom-crafter/examples/standard_test_case.groovy +0 -0
- /package/{.agents/skills/01-web-automation → automation/shared-utils}/clean-pom-crafter/references/pom_patterns.groovy +0 -0
- /package/{.agents/skills/01-web-automation → automation/shared-utils}/e2e-testing-patterns/examples/playwright_sharded_config.ts +0 -0
- /package/{.agents/skills/01-web-automation → automation/shared-utils}/e2e-testing-patterns/resources/implementation-playbook.md +0 -0
- /package/{.agents/skills/01-web-automation → automation/shared-utils}/lam-script-generator/examples/user_story_sample.md +0 -0
- /package/{.agents/skills/03-api-and-contracts → requirement/spec-review}/api-spec-to-e2e-matrix/SKILL.md +0 -0
- /package/{.agents/skills/03-api-and-contracts → requirement/spec-review}/api-spec-to-e2e-matrix/examples/sample_openapi.json +0 -0
- /package/{.agents/skills/02-test-design → requirement/spec-review}/architecture-decision-records/SKILL.md +0 -0
- /package/{.agents/skills/02-test-design → requirement/spec-review}/architecture-decision-records/examples/0001-record-architecture-decisions.md +0 -0
- /package/{.agents/skills/02-test-design → requirement/spec-review}/review-spec/examples/pocock_grilling_decision_rubric.md +0 -0
- /package/{.agents/skills/02-test-design → requirement/spec-review}/review-spec/examples/sample_spec_review.md +0 -0
- /package/{.agents → requirement}/workflows/analyze_requirement_document.md +0 -0
- /package/{.agents → requirement}/workflows/generate_requirements_from_website.md +0 -0
- /package/{.agents/rules → standards/coding-standards}/author-integrity-guard.md +0 -0
- /package/{.agents/skills/08-code-quality → standards/coding-standards}/code-refactoring-refactor-clean/SKILL.md +0 -0
- /package/{.agents/skills/08-code-quality → standards/coding-standards}/code-refactoring-refactor-clean/examples/refactored_page_sample.groovy +0 -0
- /package/{.agents/skills/08-code-quality → standards/coding-standards}/code-refactoring-refactor-clean/examples/refactored_playwright_pom.ts +0 -0
- /package/{.agents/skills/08-code-quality → standards/coding-standards}/code-refactoring-refactor-clean/resources/implementation-playbook.md +0 -0
- /package/{.agents/rules → standards/coding-standards}/communication-style.md +0 -0
- /package/{.agents/rules → standards/coding-standards}/git-push-protocol.md +0 -0
- /package/{.agents/skills/08-code-quality → standards/coding-standards}/graphify-impact-analyzer/SKILL.md +0 -0
- /package/{.agents/skills/08-code-quality → standards/coding-standards}/graphify-impact-analyzer/examples/dependency_blast_radius.json +0 -0
- /package/{.agents/skills/08-code-quality → standards/coding-standards}/lean-code-ladder/examples/ladder_walkthrough.md +0 -0
- /package/{.agents/skills/08-code-quality → standards/coding-standards}/repo-quality-audit/SKILL.md +0 -0
- /package/{.agents/skills/08-code-quality → standards/coding-standards}/repo-quality-audit/examples/sample_audit_walkthrough.md +0 -0
- /package/{.agents/skills/08-code-quality → standards/coding-standards}/repo-quality-audit/examples/sample_repo_audit.md +0 -0
- /package/{.agents/skills/08-code-quality → standards/coding-standards}/senior-code-review/SKILL.md +0 -0
- /package/{.agents/skills/08-code-quality → standards/coding-standards}/senior-code-review/examples/sample_review_walkthrough.md +0 -0
- /package/{.agents/skills/08-code-quality → standards/coding-standards}/surgical-coding-principles/examples/surgical_vs_scope_creep.md +0 -0
- /package/{.agents/skills/08-code-quality → standards/coding-standards}/technical-debt-harvester/SKILL.md +0 -0
- /package/{.agents/skills/08-code-quality → standards/coding-standards}/technical-debt-harvester/examples/sample_debt_ledger.md +0 -0
- /package/{.agents/skills/08-code-quality → standards/coding-standards}/technical-debt-harvester/examples/sample_ledger_output.md +0 -0
- /package/{.agents/rules → standards/context-ai}/context-engineering.md +0 -0
- /package/{.agents/skills/11-ai-and-prompting → standards/context-ai}/meta-prompt-crafter/SKILL.md +0 -0
- /package/{.agents/skills/11-ai-and-prompting → standards/context-ai}/meta-prompt-crafter/examples/qa_prompt_template.md +0 -0
- /package/{.agents/skills/11-ai-and-prompting → standards/context-ai}/mipro-prompt-compiler/examples/compiled_module_sample.json +0 -0
- /package/{.agents/skills/11-ai-and-prompting → standards/context-ai}/mipro-prompt-compiler/examples/dspy_assertion_optimizer_case.md +0 -0
- /package/{.agents/skills/11-ai-and-prompting → standards/context-ai}/qa-aci-interface/SKILL.md +0 -0
- /package/{.agents/skills/11-ai-and-prompting → standards/context-ai}/qa-aci-interface/examples/surgical_code_window_walkthrough.md +0 -0
- /package/{.agents/skills/11-ai-and-prompting → standards/context-ai}/qa-aci-interface/examples/surgical_edit_sample.json +0 -0
- /package/{.agents/skills/11-ai-and-prompting → standards/context-ai}/token-and-context-optimizer/SKILL.md +0 -0
- /package/{.agents/skills/11-ai-and-prompting → standards/context-ai}/token-and-context-optimizer/examples/distilled_log_sample.txt +0 -0
- /package/{.agents/skills/11-ai-and-prompting → standards/context-ai}/token-and-context-optimizer/references/context_engineering_guide.md +0 -0
- /package/{.agents/rules → standards/qa-process}/graphify.md +0 -0
- /package/{.agents/rules → standards/qa-process}/lessons-learned.md +0 -0
- /package/{.agents/rules → standards/qa-process}/skill-precedence.md +0 -0
- /package/{.agents/rules → standards/quality-gates}/systematic-debugging.md +0 -0
- /package/{.agents/templates → templates/test-report}/completion_report.md +0 -0
- /package/{.agents/skills/02-test-design → test-design/bdd-gherkin}/gherkin-to-suite/SKILL.md +0 -0
- /package/{.agents/skills/02-test-design → test-design/bdd-gherkin}/gherkin-to-suite/examples/checkout.feature +0 -0
- /package/{.agents/skills/02-test-design → test-design/bdd-gherkin}/gherkin-to-suite/examples/suite_report_sample.json +0 -0
- /package/{.agents → test-design/bdd-gherkin}/workflows/generate_cross_module_test_plan.md +0 -0
- /package/{.agents → test-design/bdd-gherkin}/workflows/generate_testcases_from_requirements.md +0 -0
- /package/{.agents/skills/02-test-design → test-design/exploratory}/tot-exploratory-testing/SKILL.md +0 -0
- /package/{.agents/skills/02-test-design → test-design/exploratory}/tot-exploratory-testing/examples/payment_gateway_tot_tree.json +0 -0
- /package/{.agents/skills/02-test-design → test-design/exploratory}/tot-exploratory-testing/examples/tot_graph_sample.json +0 -0
- /package/{.agents/skills/02-test-design → test-design/manual-rbt}/rbt-manual-testing/examples/checkout_rbt_manual_suite.md +0 -0
- /package/{.agents/skills/02-test-design → test-design/manual-rbt}/rbt-manual-testing/references/field_validation_checklist.md +0 -0
- /package/{.agents → test-design/manual-rbt}/workflows/export_testcases_excel.md +0 -0
- /package/{.agents → test-design/manual-rbt}/workflows/generate_manual_testcases_rbt.md +0 -0
- /package/{.agents/skills/07-data-and-state → test-design/test-data}/db-state-verifier-and-seeder/SKILL.md +0 -0
- /package/{.agents/skills/07-data-and-state → test-design/test-data}/db-state-verifier-and-seeder/examples/db_fixture_seed.json +0 -0
- /package/{.agents/skills/07-data-and-state → test-design/test-data}/file-and-report-validator/examples/sample_export_schema.json +0 -0
- /package/{.agents/skills/07-data-and-state → test-design/test-data}/synthetic-test-data-generator/examples/idempotent_data_recipe.json +0 -0
- /package/{.agents → test-design/test-data}/workflows/generate_combinatorial_test_data.md +0 -0
- /package/{.agents/skills/03-api-and-contracts → testing/api-contracts}/api-network-mocking-interception/SKILL.md +0 -0
- /package/{.agents/skills/03-api-and-contracts → testing/api-contracts}/api-network-mocking-interception/examples/network_route_stub.json +0 -0
- /package/{.agents/skills/03-api-and-contracts → testing/api-contracts}/consumer-driven-contract-testing/SKILL.md +0 -0
- /package/{.agents/skills/03-api-and-contracts → testing/api-contracts}/consumer-driven-contract-testing/examples/order_service_pact.json +0 -0
- /package/{.agents/skills/06-healing-and-resilience → testing/healing-resilience}/agentless-qa-repair/examples/healing_report_sample.json +0 -0
- /package/{.agents/skills/06-healing-and-resilience → testing/healing-resilience}/agentless-qa-repair/examples/three_phase_repair_workflow.md +0 -0
- /package/{.agents/skills/06-healing-and-resilience → testing/healing-resilience}/antigravity-test-sentinel/SKILL.md +0 -0
- /package/{.agents/skills/06-healing-and-resilience → testing/healing-resilience}/antigravity-test-sentinel/examples/dod_gate_policy.json +0 -0
- /package/{.agents/skills/06-healing-and-resilience → testing/healing-resilience}/antigravity-test-sentinel/examples/sprint_dod_enforcement.md +0 -0
- /package/{.agents/skills/06-healing-and-resilience → testing/healing-resilience}/failure-autopsy-and-healing/examples/closed_loop_healing_walkthrough.md +0 -0
- /package/{.agents/skills/06-healing-and-resilience → testing/healing-resilience}/failure-autopsy-and-healing/examples/sample_junit_failure.xml +0 -0
- /package/{.agents/skills/06-healing-and-resilience → testing/healing-resilience}/failure-autopsy-and-healing/references/healing_matrix.md +0 -0
- /package/{.agents/skills/06-healing-and-resilience → testing/healing-resilience}/failure-autopsy-and-healing/references/qa_investigation_protocol.md +0 -0
- /package/{.agents/skills/06-healing-and-resilience → testing/healing-resilience}/failure-autopsy-and-healing/scripts/triage_failures.sh +0 -0
- /package/{.agents/skills/06-healing-and-resilience → testing/healing-resilience}/flaky-test-sentinel-and-quarantine/SKILL.md +0 -0
- /package/{.agents/skills/06-healing-and-resilience → testing/healing-resilience}/flaky-test-sentinel-and-quarantine/examples/quarantine_registry.json +0 -0
- /package/{.agents/skills/06-healing-and-resilience → testing/healing-resilience}/self-healing-locator/examples/aom_diff_sample.json +0 -0
- /package/{.agents/skills/06-healing-and-resilience → testing/healing-resilience}/systematic-debugging/SKILL.md +0 -0
- /package/{.agents/skills/06-healing-and-resilience → testing/healing-resilience}/systematic-debugging/examples/four_phase_case.md +0 -0
- /package/{.agents/skills/06-healing-and-resilience → testing/healing-resilience}/test-driven-development/SKILL.md +0 -0
- /package/{.agents/skills/06-healing-and-resilience → testing/healing-resilience}/test-driven-development/examples/red_green_refactor_qa.md +0 -0
- /package/{.agents/skills/06-healing-and-resilience → testing/healing-resilience}/verification-before-completion/SKILL.md +0 -0
- /package/{.agents/skills/06-healing-and-resilience → testing/healing-resilience}/verification-before-completion/examples/evidence_log_example.md +0 -0
- /package/{.agents/skills/06-healing-and-resilience → testing/healing-resilience}/visual-regression-sentinel/SKILL.md +0 -0
- /package/{.agents/skills/06-healing-and-resilience → testing/healing-resilience}/visual-regression-sentinel/examples/visual_mask_rules.json +0 -0
- /package/{.agents/skills/04-performance-and-load → testing/performance}/backend-load-and-stress-patterns/examples/k6_load_test_sample.js +0 -0
- /package/{.agents/skills/04-performance-and-load → testing/performance}/backend-load-and-stress-patterns/references/k6_load_template.js +0 -0
- /package/{.agents/skills/04-performance-and-load → testing/performance}/lean-impact-scoreboard/SKILL.md +0 -0
- /package/{.agents/skills/04-performance-and-load → testing/performance}/lean-impact-scoreboard/examples/sample_benchmark_gain.md +0 -0
- /package/{.agents/skills/04-performance-and-load → testing/performance}/lean-impact-scoreboard/examples/sample_scoreboard.md +0 -0
- /package/{.agents/skills/04-performance-and-load → testing/performance}/web-vitals-and-performance-profiler/examples/cwv_budget_policy.json +0 -0
- /package/{.agents/skills/04-performance-and-load → testing/performance}/web-vitals-and-performance-profiler/examples/playwright_cwv_profiler_fixture.ts +0 -0
- /package/{.agents/skills/05-security-and-compliance → testing/security-a11y}/security-dast-sentinel/examples/dast_scan_config.json +0 -0
- /package/{.agents/skills/05-security-and-compliance → testing/security-a11y}/security-dast-sentinel/references/AI-AND-LLM.md +0 -0
- /package/{.agents/skills/05-security-and-compliance → testing/security-a11y}/security-dast-sentinel/references/ATTACK-CLASSES.md +0 -0
- /package/{.agents/skills/05-security-and-compliance → testing/security-a11y}/security-dast-sentinel/references/dast_recipes.ts +0 -0
- /package/{.agents/skills/05-security-and-compliance → testing/security-a11y}/skillspector-security-guard/SKILL.md +0 -0
- /package/{.agents/skills/05-security-and-compliance → testing/security-a11y}/skillspector-security-guard/examples/audit_threat_vectors_report.json +0 -0
- /package/{.agents/skills/05-security-and-compliance → testing/security-a11y}/skillspector-security-guard/examples/safe_mcp_config.json +0 -0
- /package/{.agents/skills/05-security-and-compliance → testing/security-a11y}/wcag-audit-patterns/examples/wcag_compliance_checklist.json +0 -0
- /package/{.agents/skills/05-security-and-compliance → testing/security-a11y}/wcag-audit-patterns/resources/implementation-playbook.md +0 -0
- /package/{.agents/skills/01-web-automation → tools/benchmarks-tests}/web-agent-gym-benchmarking/SKILL.md +0 -0
- /package/{.agents/skills/01-web-automation → tools/benchmarks-tests}/web-agent-gym-benchmarking/examples/benchmark_report_sample.json +0 -0
- /package/{.agents/skills/01-web-automation → tools/benchmarks-tests}/web-agent-gym-benchmarking/examples/enterprise_tasks_suite.json +0 -0
|
@@ -0,0 +1,1038 @@
|
|
|
1
|
+
# COMPREHENSIVE ENTERPRISE QUALITY & ARCHITECTURAL AUDIT REPORT
|
|
2
|
+
# SYNAPSE HUB ENTERPRISE QA PLATFORM
|
|
3
|
+
|
|
4
|
+
> 🛡️ **Synapse Hub Active:** Clean POM / Flat Architecture
|
|
5
|
+
|
|
6
|
+
**Target Repository:** Synapse Hub (`https://github.com/TeeVee98/Synapse-Hub`)
|
|
7
|
+
**Audit Integrity Mode:** Benchmark (Strictly Read-Only, AST Analyzed, Empirical Verification)
|
|
8
|
+
**Lead Auditor:** Lead Master Technical Writer & Systems Auditor
|
|
9
|
+
**Date of Audit:** 2026-11-14
|
|
10
|
+
**Audit Scope:** 360-Degree Evaluation across 7 Core Enterprise Engineering Pillars
|
|
11
|
+
|
|
12
|
+
---
|
|
13
|
+
|
|
14
|
+
## 1. EXECUTIVE SCORECARD
|
|
15
|
+
|
|
16
|
+
### 1.1 Executive Summary
|
|
17
|
+
An exhaustive, adversarial, 360-degree systems audit was conducted on Synapse Hub across its core architecture, token optimization engines, multi-agent coordination frameworks, self-healing mechanisms, security controls, and release pipelines. Synapse Hub represents an exceptionally sophisticated, production-grade enterprise QA automation platform. It distinguishes itself through mathematically rigorous context optimization, deterministic finite-state coordination, closed-loop AST self-healing, transactional write-ahead logging rollback, and unified polyglot release automation.
|
|
18
|
+
|
|
19
|
+
The audit verified **120+ automated regression and unit tests**, executed all **10 Quality Gates** of the Definition of Done (DoD) engine, and validated operational behavior across real-world polyglot reference test suites (Playwright TypeScript, Katalon Groovy, Cypress JavaScript, Pytest Python, and Selenium Python).
|
|
20
|
+
|
|
21
|
+
While the foundational architecture achieves world-class ratings, adversarial probing identified **13 classified findings** (6 P1 High Severity, 7 P2 Medium Severity, 0 P0 Critical Vulnerabilities). These findings primarily center around static analyzer boundary conditions, coarse baseline suppression granularity, and edge-case CLI ergonomics. With targeted hotfixes applied to the 6 P1 findings, Synapse Hub qualifies for full enterprise mission-critical deployment.
|
|
22
|
+
|
|
23
|
+
---
|
|
24
|
+
|
|
25
|
+
### 1.2 Pillar Evaluation Scorecard
|
|
26
|
+
|
|
27
|
+
| # | Pillar Dimension | Score (1-10) | Status | Key Empirical Metrics & Core Rationale |
|
|
28
|
+
|---|---|:---:|:---:|---|
|
|
29
|
+
| **1** | **Token Engineering & Context Optimization** | **10.0 / 10** | **SUPERIOR** | **69.1% (Tier 1) & 79.2% (Tier 2) context reduction**; dynamic disk enumeration (13 rules = 8,301 tokens, 53 skills = 55,996 tokens, unrouted baseline = 64,297 tokens); exact BPE tokenizer (`cl100k_base`, `is_exact=True`); truth anchor enforces 0.0% savings on unclassified workspaces; byte-for-byte benchmark drift sentinel; hard-capped log distillation. |
|
|
30
|
+
| **2** | **Clean POM Flat Architecture & Test Structure** | **8.5 / 10** | **STRONG** | 100% locator encapsulation in reference Page Classes; zero static sleeps in reference suites; dynamic test entity randomization (`uniqueCode("AUTO_")`). Scored down due to Gate 7 sub-3000ms static sleep pass-through (**F-P2-01**), Gate 3 regex blindspot for Playwright semantic locators (**F-P2-02**), and missing teardown in Selenium order test (**F-P2-03**). |
|
|
31
|
+
| **3** | **Multi-Agent Swarm & Finite State Machine (FSM)** | **9.5 / 10** | **SUPERIOR** | Single Front-Door Orchestrator with Lean Coordination; 5-Stage deterministic FSM (`tools/fsm_enforcer.py`); active runtime middleware (`tools/fsm_middleware.py`) blocking unauthorized tool calls; zero trigger collisions across 8 specialized subagents; clear skill precedence hierarchy. Minor: in-context plugin suppression (**F-P3-01**). |
|
|
32
|
+
| **4** | **Autonomous Self-Healing & AST Mutation Testing** | **10.0 / 10** | **SUPERIOR** | Noah Shinn Reflexion (K=3) with AOM pruning (< 500 tokens / 1,800 chars); negative feedback cache (`agent-qa heal record-failed`) eliminating LLM blind retry loops; 2PC Write-Ahead Logging (WAL) reversible sandbox with `SIGINT`/`SIGTERM`/`atexit` traps; micro-mutator with 24+ operator families and AST syntax error discard (Mutation Kill Score: 86.8%–100.0%). |
|
|
33
|
+
| **5** | **Security, Hygiene & Author Attribution Guard** | **8.0 / 10** | **ADEQUATE** | 100% UTF-8 No BOM across all 660 candidate files; zero hardcoded plaintext credentials; cryptographic SHA-256 author integrity guard for Thế Vị (@TeeVee98). Scored down due to Gate 2 `\b/` word-boundary regex flaw masking 7 tracked absolute paths (**SEC-01**), pre-commit guard skipping `.py` files (**SEC-02**), and 46.8% missing internal source SPDX headers (**SEC-03**). |
|
|
34
|
+
| **6** | **10 Quality Gates Definition of Done (DoD)** | **8.0 / 10** | **ADEQUATE** | Automated 10-Gate verification runner (`tools/verify_all.py`); static AST syntax verification; SLA performance enforcement (P95 <= 500ms). Scored down due to baseline carte blanche whole-file suppression in Gate 7 (**QG-01**), 78 occurrences of dummy bypass assignments `_ = err` (**QG-02**), and corrupted JSON output in CLI telemetry (**QG-03**). |
|
|
35
|
+
| **7** | **Developer Experience & Zero-Command Invariant** | **9.5 / 10** | **SUPERIOR** | 19 autonomous intent-mapped tools without CLI syntax memorization; universal version bumper synchronizing 20 distinct targets atomically with rollback and dry-run; 5-step automated release publisher with GitHub Actions OIDC Trusted Publishing and NPM verification. Scored down slightly due to CLI argument crash in `project_state.py` (**F-P7-01**). |
|
|
36
|
+
|
|
37
|
+
---
|
|
38
|
+
|
|
39
|
+
### 1.3 Overall Repository Score Calculation
|
|
40
|
+
The overall repository health score is calculated as the unweighted arithmetic mean of the 7 audited pillars:
|
|
41
|
+
|
|
42
|
+
$$\text{Overall Repository Score} = \frac{10.0 + 8.5 + 9.5 + 10.0 + 8.0 + 8.0 + 9.5}{7} = \frac{63.5}{7} \approx \mathbf{9.07 \text{ / } 10 \quad (90.7\% \text{ — Grade: A-})}$$
|
|
43
|
+
|
|
44
|
+
**Enterprise Readiness Rating:** **CONDITIONAL ENTERPRISE PRODUCTION READY**
|
|
45
|
+
*The platform core is exceptionally mature. Production deployment is authorized subject to applying the recommended hotfixes for the 6 P1 findings.*
|
|
46
|
+
|
|
47
|
+
---
|
|
48
|
+
|
|
49
|
+
## 2. TOP 5 COMPETITIVE ADVANTAGES
|
|
50
|
+
|
|
51
|
+
Synapse Hub possesses architectural innovations that place it significantly ahead of standard industry test automation frameworks and AI agent harnesses. Below is a deep architectural analysis of its 5 most impressive, battle-tested capabilities.
|
|
52
|
+
|
|
53
|
+
---
|
|
54
|
+
|
|
55
|
+
### 2.1 Deterministic 5-Stage Finite State Machine (FSM) Enforcer & Runtime Middleware Interception
|
|
56
|
+
Most multi-agent systems rely exclusively on prompt-based instructions to govern agent behavior. Synapse Hub couples conversational guidelines with a deterministic, programmatic state machine and runtime middleware interception.
|
|
57
|
+
|
|
58
|
+
```
|
|
59
|
+
┌────────────────┐
|
|
60
|
+
│ DISCOVERY │ ◄── live-explorer / test-designer
|
|
61
|
+
└───────┬────────┘ (Requires selectors, DOM hash, AOM nodes, test matrix)
|
|
62
|
+
▼
|
|
63
|
+
┌────────────────┐
|
|
64
|
+
│ CODE_SYNTHESIS │ ◄── pom-crafter / api-engineer
|
|
65
|
+
└───────┬────────┘ (Requires page classes, clean POM scripts, API matrix)
|
|
66
|
+
▼
|
|
67
|
+
┌────────────────┐
|
|
68
|
+
│ HEADLESS_EXEC │ ◄── test-runner / perf-engineer / api-engineer
|
|
69
|
+
└───────┬────────┘ (Requires non-null integer exit_code)
|
|
70
|
+
▼
|
|
71
|
+
┌────────────────┐
|
|
72
|
+
│ AUTOPSY_HEAL │ ◄── failure-analyst
|
|
73
|
+
└───────┬────────┘ (Requires root_cause, patch, or PASSED_NO_HEAL_NEEDED)
|
|
74
|
+
▼
|
|
75
|
+
┌────────────────┐
|
|
76
|
+
│ ACCEPTANCE_DOD │ ◄── quality-gatekeeper
|
|
77
|
+
└────────────────┘ (Requires all 10 Gates PASS)
|
|
78
|
+
```
|
|
79
|
+
|
|
80
|
+
1. **State Machine Core (`tools/fsm_enforcer.py`)**:
|
|
81
|
+
- Manages state through `STAGES_ORDER = ["DISCOVERY", "CODE_SYNTHESIS", "HEADLESS_EXEC", "AUTOPSY_HEAL", "ACCEPTANCE_DOD"]`.
|
|
82
|
+
- Maps each stage strictly to authorized actors (`STAGE_ACTORS`), preventing out-of-order execution (e.g. an agent cannot jump to `CODE_SYNTHESIS` without completing `DISCOVERY`).
|
|
83
|
+
- Validates formal data schemas (`STAGE_EVIDENCE_SCHEMAS`). For example, `HEADLESS_EXEC` rejects transitions unless a valid integer `exit_code` is provided; `AUTOPSY_HEAL` mandates a structured root cause and patch attempt or explicit pass waiver (`PASSED_NO_HEAL_NEEDED`).
|
|
84
|
+
2. **Runtime Tool Interception Middleware (`tools/fsm_middleware.py`)**:
|
|
85
|
+
- Intercepts subagent tool invocations before execution.
|
|
86
|
+
- Code mutation tools (`write_to_file`, `replace_file_content`, `edit_file`) are blocked with HTTP/Runtime exceptions if Stage 1 (`DISCOVERY`) is incomplete.
|
|
87
|
+
- Test execution tools (`run_test`, `run_playwright`, `run_pytest`) are blocked if Stage 2 (`CODE_SYNTHESIS`) is incomplete.
|
|
88
|
+
- Task completion triggers are blocked until Stage 5 (`ACCEPTANCE_DOD`) verifies the 10 Quality Gates.
|
|
89
|
+
- Read-only tools (`view_file`, `list_dir`, `grep_search`) pass freely without state mutation.
|
|
90
|
+
3. **Live Data Contracts (`tools/contract_validator.py`)**:
|
|
91
|
+
- Validates JSON Data Contracts exchanged between `live-explorer` and `pom-crafter`.
|
|
92
|
+
- Mandates semantic node attributes (`role`, `name`, `testId`) and enforces interaction latency below 200ms.
|
|
93
|
+
|
|
94
|
+
---
|
|
95
|
+
|
|
96
|
+
### 2.2 Dual-Tier Context Hydration & Adversarial Truth Anchor
|
|
97
|
+
Synapse Hub solves context window bloat and speculative prompt inflation through a mathematically verified dual-tier hydration architecture.
|
|
98
|
+
|
|
99
|
+
```
|
|
100
|
+
Full Disk Catalog Baseline
|
|
101
|
+
[ 13 Rules: 8,301 Tokens | 53 Skills: 55,996 Tokens ]
|
|
102
|
+
Total: 64,297 Tokens
|
|
103
|
+
│
|
|
104
|
+
┌────────────────────┴────────────────────┐
|
|
105
|
+
▼ ▼
|
|
106
|
+
Tier 1: Orchestrator Tier 2: Subagent
|
|
107
|
+
Catalog Metadata Routing Full Workload Hydration
|
|
108
|
+
────────────────────────────── ──────────────────────────────
|
|
109
|
+
Baseline : 11,375 Tokens Baseline : 64,297 Tokens
|
|
110
|
+
Routed : 3,517 Tokens Routed : 13,400 Tokens
|
|
111
|
+
Reduction: 69.1% SAVINGS Reduction: 79.2% SAVINGS
|
|
112
|
+
Payload : Catalog Summaries Payload : Full Methodologies
|
|
113
|
+
```
|
|
114
|
+
|
|
115
|
+
1. **Physical Disk Enumeration (`tools/core/token_estimator.py`)**:
|
|
116
|
+
- Replaces theoretical multipliers with live filesystem walks. Enumerate 13 rule files (`.agents/rules/*.md`) and 53 skill files (`.agents/skills/*/*/SKILL.md`).
|
|
117
|
+
- Empirically measured: 13 rules = 8,301 tokens (32,351 bytes); 53 skills = 55,996 tokens (232,833 bytes); full unrouted baseline = 64,297 tokens (265,184 bytes); summary baseline = 11,375 tokens.
|
|
118
|
+
2. **Dual-Tier Hydration Contracts**:
|
|
119
|
+
- **Tier 1 (Catalog Summary)**: Hydrates the Single Front-Door Orchestrator with skill headers and summaries (~3,517 tokens, achieving a **69.1% context reduction**).
|
|
120
|
+
- **Tier 2 (Full Hydration)**: Hydrates specialized subagents with complete behavioral rulebooks and method implementations (~13,400 tokens, achieving a **79.2% context reduction**).
|
|
121
|
+
- Disk I/O avoided is tracked separately (208,791 bytes avoided, **78.7% I/O reduction**), maintaining clean separation between prompt tokens and storage reads.
|
|
122
|
+
3. **Adversarial Truth Anchor**:
|
|
123
|
+
- When executed against an unclassified directory (e.g. `/tmp` or an empty workspace), `tools/rules_router.py` returns `status: UNCLASSIFIED_WORKSPACE`, `confidence_score: 0.0`, `tokens_saved: 0`, and `reduction_percentage: 0.0%`.
|
|
124
|
+
- Unlike naive AI agents that report simulated 80% savings on empty directories, Synapse Hub refuses to claim token reductions when no workload specialization is detected.
|
|
125
|
+
4. **Byte-for-Byte Benchmark Drift Sentinel (`tests/test_benchmark_context.py`)**:
|
|
126
|
+
- The test suite executes `tools/benchmark_swarm_context.py` and performs byte-for-byte matching against `benchmarks/CONTEXT_TOKEN_BENCHMARK.md`. Any undocumented modification to token metrics breaks CI/CD automatically.
|
|
127
|
+
|
|
128
|
+
---
|
|
129
|
+
|
|
130
|
+
### 2.3 Noah Shinn Reflexion (K=3) Closed-Loop Self-Healing with AOM Pruning & Negative Feedback Cache
|
|
131
|
+
Self-healing in Synapse Hub incorporates episodic reinforcement learning principles inspired by the Noah Shinn Reflexion architecture, overcoming the fragility of single-shot heuristic regex replacements.
|
|
132
|
+
|
|
133
|
+
```
|
|
134
|
+
┌────────────────────────────────────────────────────────────────────────┐
|
|
135
|
+
│ Failing Test Run Detected │
|
|
136
|
+
└───────────────────────────────────┬────────────────────────────────────┘
|
|
137
|
+
▼
|
|
138
|
+
Extract Raw DOM / CDP Accessibility Tree
|
|
139
|
+
│
|
|
140
|
+
▼
|
|
141
|
+
Prune AOM to Compact Subtree (< 500 Tokens)
|
|
142
|
+
(Retains: role, name, testId, disabled, state)
|
|
143
|
+
│
|
|
144
|
+
▼
|
|
145
|
+
┌─────────────► Closed-Loop Reflexion Loop (K=3)
|
|
146
|
+
│ │
|
|
147
|
+
│ Trial t in [1..3] ▼
|
|
148
|
+
│ Synthesize Verbal Reflection & Root Cause
|
|
149
|
+
│ (e.g. ACCESSIBLE_NAME_RESOLVED, ELEMENT_DISABLED)
|
|
150
|
+
│ │
|
|
151
|
+
│ ▼
|
|
152
|
+
│ Generate Candidate Clean POM Patch
|
|
153
|
+
│ (e.g. page.getByRole('button', { name: 'Submit' }))
|
|
154
|
+
│ │
|
|
155
|
+
│ ▼
|
|
156
|
+
│ Execute Verification in Isolated Sandbox
|
|
157
|
+
│ │
|
|
158
|
+
├─── [VERIFY: FAIL] ───────┤
|
|
159
|
+
│ Accumulate memory; │
|
|
160
|
+
│ t = t + 1 ▼
|
|
161
|
+
│ [VERIFY: PASS]
|
|
162
|
+
│ │
|
|
163
|
+
▼ ▼
|
|
164
|
+
[Exhausted K=3] Record Positive Memory
|
|
165
|
+
│ Commit Patch to Page Class
|
|
166
|
+
▼
|
|
167
|
+
Record Negative Memory
|
|
168
|
+
SQLite `failed_fixes` Table
|
|
169
|
+
(agent-qa heal record-failed)
|
|
170
|
+
Banned from future retry loops!
|
|
171
|
+
```
|
|
172
|
+
|
|
173
|
+
1. **AOM Compact Pruning (`tools/auto_heal_reflexion.py:61-141`)**:
|
|
174
|
+
- Raw DOM trees often consume 50,000+ tokens. The pruner recursively extracts only accessibility-salient attributes (`role`, `name`, `aria-label`, `data-testid`, `disabled`, `expanded`, `checked`).
|
|
175
|
+
- Caps depth $\le 4$ and children $\le 12$, bounding the total representation to $< 1,800$ characters ($< 500$ tokens).
|
|
176
|
+
2. **K=3 Multi-Trial Reflection Loop**:
|
|
177
|
+
- Evaluates up to 3 candidate patches ($K=3$). If Trial 1 fails, Trial 2 receives the failing error, the prior attempt, and a verbal reflection analyzing why the prior attempt was invalid.
|
|
178
|
+
- Diagnoses semantic root causes: `ACCESSIBLE_NAME_OR_ROLE_RESOLVED`, `TARGET_ELEMENT_DISABLED`, `TARGET_ELEMENT_DETACHED`.
|
|
179
|
+
3. **Negative Feedback Cache (`agent-qa heal record-failed`)**:
|
|
180
|
+
- When all 3 trials are exhausted without success, the failure signature, patch code, and root cause are persisted to SQLite (`tools/cache_resolver.py:566-581`).
|
|
181
|
+
- Future healing attempts query this negative memory blacklist, preventing AI agents from repeating known failing patches across subsequent turns.
|
|
182
|
+
|
|
183
|
+
---
|
|
184
|
+
|
|
185
|
+
### 2.4 Reversible Sandbox Two-Phase Commit (2PC) Write-Ahead Logging (WAL) Engine
|
|
186
|
+
Test automation and automated self-healing scripts frequently mutate files, insert database rows, and modify environment state. An unhandled crash or abort (`SIGINT`) often leaves orphan test data and corrupts filesystems. Synapse Hub introduces a database-grade transactional engine for filesystem mutations.
|
|
187
|
+
|
|
188
|
+
```
|
|
189
|
+
Test / Self-Healing Mutation Initiated
|
|
190
|
+
│
|
|
191
|
+
▼
|
|
192
|
+
Phase 1: Write-Ahead Journaling (.synapse/undo_wal.jsonl)
|
|
193
|
+
(Logs: action_type, target_resource, undo_command, lock)
|
|
194
|
+
│
|
|
195
|
+
▼
|
|
196
|
+
Phase 2: Execute Mutation on Target Resource
|
|
197
|
+
│
|
|
198
|
+
▼
|
|
199
|
+
┌───────────────────────┴───────────────────────┐
|
|
200
|
+
▼ ▼
|
|
201
|
+
[SUCCESS] [FAIL / ABORT]
|
|
202
|
+
Test Completes (SIGINT, SIGTERM, crash)
|
|
203
|
+
│ │
|
|
204
|
+
▼ ▼
|
|
205
|
+
LIFO Stack Rollback Emergency Trap Recovery
|
|
206
|
+
Reverse Chronological Order Read WAL, Execute Reverse Undo
|
|
207
|
+
│ │
|
|
208
|
+
▼ ▼
|
|
209
|
+
All Actions Succeeded? Preserve Failed Actions
|
|
210
|
+
│ (status: FAILED in WAL)
|
|
211
|
+
┌───────┴───────┐ │
|
|
212
|
+
YES NO ▼
|
|
213
|
+
│ │ Survives System Restart!
|
|
214
|
+
▼ ▼
|
|
215
|
+
Purge WAL Keep Failed WAL
|
|
216
|
+
```
|
|
217
|
+
|
|
218
|
+
1. **LIFO Undo Stack & WAL Persistence (`tools/reversible_sandbox.py`)**:
|
|
219
|
+
- Maintains an in-memory stack of `ReversibleAction` records paired with atomic disk logging to `.synapse/undo_wal.jsonl` using process file locking.
|
|
220
|
+
- Supports native undo primitives: `FILE_DELETE:`, `DIR_DELETE:`, `NOOP`, or arbitrary subshell shell commands.
|
|
221
|
+
2. **Crash & Signal Resiliency**:
|
|
222
|
+
- Registers OS signal handlers for `SIGINT` and `SIGTERM` as well as an `atexit` hook. If an engineer cancels execution (`Ctrl+C`), the sandbox intercepts the signal and rolls back all mutations in LIFO order before terminating with `128 + signum`.
|
|
223
|
+
3. **Two-Phase Commit Guarantees**:
|
|
224
|
+
- Incremental WAL management clears successfully unwound actions while marking failed rollbacks as `status: FAILED`.
|
|
225
|
+
- The WAL is purged if and only if 100% of recorded actions are cleanly reversed. If an undo fails, the log remains on disk, allowing offline recovery via `recover_from_wal()`.
|
|
226
|
+
|
|
227
|
+
---
|
|
228
|
+
|
|
229
|
+
### 2.5 Polyglot Universal Version Synchronizer & Autonomous OIDC Hub Publisher
|
|
230
|
+
Maintaining semantic version synchronization across polyglot ecosystems (Node.js, Python, VS Code, documentation, manifests) is a notorious source of release drift. Synapse Hub features an automated release orchestrator.
|
|
231
|
+
|
|
232
|
+
1. **Polyglot Synchronization Engine (`tools/bump_version.py`)**:
|
|
233
|
+
- Synchronizes semantic version strings across **20 distinct project targets** in a single atomic pass:
|
|
234
|
+
- Package manifests: `package.json`, `plugin.json`, `editors/vscode/package.json`, `pyproject.toml`, `agentskills.json`.
|
|
235
|
+
- Configuration & catalogs: `skills_manifest.json`, `synapse.config.json`.
|
|
236
|
+
- Code & CLI headers: `tools/cli.py`, `tools/mcp_server.py`, `tools/browser_helpers.js`, `tools/mcp_excel_server.py`, `tools/mcp_docx_server.py`, `tools/__init__.py`, `tools/ui_server.py`, `tools/welcome_page.py`, `tools/dashboard_generator.py`, `tools/update_sentinel.py`.
|
|
237
|
+
- Documentation: `docs/ARCHITECTURE.md`, `docs/ARCHITECTURE.en.md`, `README.md`, `README.en.md`, `SECURITY.md`, `CHANGELOG.md`.
|
|
238
|
+
- Includes dry-run simulation (`--dry-run`) and automatic rollback on verification failure.
|
|
239
|
+
2. **Autonomous 5-Step Release Pipeline (`tools/hub_publisher.py`)**:
|
|
240
|
+
- Step 1: Git push to `origin/main`.
|
|
241
|
+
- Step 2: Creates and pushes annotated Git tag `vX.Y.Z`.
|
|
242
|
+
- Step 3: Invokes GitHub REST API to generate release notes and publish release.
|
|
243
|
+
- Step 4: Builds and publishes GitHub Packages artifact (`@teevee98/synapse-hub`).
|
|
244
|
+
- Step 5: Publishes to NPM Registry using GitHub Actions OIDC Trusted Publishing and polls the live registry until the package is globally resolvable.
|
|
245
|
+
- Pre-flight quality gates execute all 10 DoD gates and regression suites before any remote network call is permitted.
|
|
246
|
+
|
|
247
|
+
---
|
|
248
|
+
|
|
249
|
+
## 3. CLASSIFIED FINDINGS MATRIX & DETAILED AUDIT REPORTS
|
|
250
|
+
|
|
251
|
+
### 3.1 Classified Findings Matrix
|
|
252
|
+
|
|
253
|
+
The adversarial audit identified 13 total findings: 0 P0 (Critical), 6 P1 (High), and 7 P2 (Medium).
|
|
254
|
+
|
|
255
|
+
| Finding ID | Severity | Pillar | Component | Summary of Finding | Exact Location Citation |
|
|
256
|
+
|---|:---:|:---:|---|---|---|
|
|
257
|
+
| **SEC-01** | **P1** | Pillar 5 | `tools/verify_all.py` Gate 2 | `\b/Users/` and `\b/home/` regex boundary flaw misses quoted/spaced paths; 7 tracked machine paths bypassed Gate 2. | `tools/verify_all.py:93-94`, `docs/PERFORMANCE_AUDIT_REPORT.md:6, 552` |
|
|
258
|
+
| **SEC-02** | **P1** | Pillar 5 | `tools/git_pre_commit_guard.py` | `check_forbidden_paths` skips all `.py` files, completely blinding pre-commit hooks to paths in Python code. | `tools/git_pre_commit_guard.py:99` |
|
|
259
|
+
| **QG-01** | **P1** | Pillar 6 | `tools/infer_linter.py` Gate 7 | Baseline suppression operates at whole-file scope `(f_path, d_type)`, granting carte blanche immunity to new defects. | `tools/infer_linter.py:258, 730` |
|
|
260
|
+
| **QG-02** | **P1** | Pillar 6 | `tools/` Internal Codebase | 78 occurrences of dummy bypass assignments (`_ = err`, `_ = False`) masking potential errors in tools and tests. | `tools/verify_all.py:73, 78`, `tools/infer_linter.py:226`, 27 other files |
|
|
261
|
+
| **F-P2-01** | **P1** | Pillar 2 | `tools/infer_linter.py` Gate 7 | Sub-3000ms static sleeps (`time.sleep(1)`) assigned `WARNING` instead of `CRITICAL`, passing `--strict` Gate 7. | `tools/infer_linter.py:447-449, 475-477` |
|
|
262
|
+
| **F-P2-02** | **P1** | Pillar 2 | `tools/verify_all.py` Gate 3 | Clean POM regex only matches CSS/XPaths, blind to Playwright semantic locators (`getByRole`, etc.) in test scripts. | `tools/verify_all.py:373-401` |
|
|
263
|
+
| **SEC-03** | **P2** | Pillar 5 | `tools/author_integrity.py` | 132/282 source files (46.8%) lack SPDX/author copyright headers; integrity tool only checks 4 root files. | Repository source files (`tools/`, `tests/`) |
|
|
264
|
+
| **SEC-04** | **P2** | Pillar 5 | `tools/git_pre_commit_guard.py` | Pre-commit hook lacks secret scanning (Gate 6) and author attribution pre-flight validation. | `tools/git_pre_commit_guard.py` |
|
|
265
|
+
| **QG-03** | **P2** | Pillar 6 | `tools/verify_all.py` Output | `safe_print` default argument `stream=sys.stdout` leaks progress banners to stdout during `--format json`, breaking CI. | `tools/verify_all.py:58, 851, 877, 1259` |
|
|
266
|
+
| **F-P2-03** | **P2** | Pillar 2 | `examples/selenium-suite` | Selenium reference test creates orders without teardown deletion fixture, violating AGENTS.md § 1.3. | `examples/selenium-suite/tests/test_order.py:4-10` |
|
|
267
|
+
| **F-P3-01** | **P2** | Pillar 3 | `.agents/rules/skill-precedence.md` | External plugin suppression relies on in-context LLM compliance rather than OS process/sandbox isolation. | `.agents/rules/skill-precedence.md:5-20` |
|
|
268
|
+
| **F-P7-01** | **P2** | Pillar 7 | `tools/project_state.py` | `project_state.py status` crashes with `FileNotFoundError` due to handling subcommand string as directory path. | `tools/project_state.py:169-170` |
|
|
269
|
+
| **ROUTER-01** | **P2** | Pillar 1 | `tools/rules_router.py` | Directory marker scan depth capped at `depth >= 2`, missing stack markers in deeply nested monorepo packages. | `tools/rules_router.py:306-314` |
|
|
270
|
+
|
|
271
|
+
---
|
|
272
|
+
|
|
273
|
+
### 3.2 Detailed Finding Reports (Root Cause, Risk, Reproduction & Code Remediation)
|
|
274
|
+
|
|
275
|
+
---
|
|
276
|
+
|
|
277
|
+
#### Finding SEC-01: Word Boundary Regex Flaw in Gate 2 Path Portability Checker
|
|
278
|
+
- **Severity**: **P1 (High)**
|
|
279
|
+
- **Pillar**: Pillar 5 (Security, Hygiene & Author Attribution Guard)
|
|
280
|
+
- **Exact Location**: `tools/verify_all.py:93-94`
|
|
281
|
+
- **Quoted Source Code**:
|
|
282
|
+
```python
|
|
283
|
+
# tools/verify_all.py:90-97
|
|
284
|
+
FORBIDDEN_PATH_PATTERNS = [
|
|
285
|
+
(re.compile(r'file:///[a-zA-Z]:', re.IGNORECASE), "Machine file URI (file:///...)"),
|
|
286
|
+
(re.compile(r'\b[a-zA-Z]:\\Users\\[a-zA-Z0-9_\-]+\\', re.IGNORECASE), "Windows User directory path"),
|
|
287
|
+
(re.compile(r'\b/Users/[a-zA-Z0-9_\-]+/', re.IGNORECASE), "macOS/Unix User directory path"),
|
|
288
|
+
(re.compile(r'\b/home/[a-zA-Z0-9_\-]+/', re.IGNORECASE), "Linux User home directory path"),
|
|
289
|
+
(re.compile(r'\b[a-zA-Z]:[/\\](?:agent-skills|workspace|jenkins|builds)\b', re.IGNORECASE), "Machine-specific repository path"),
|
|
290
|
+
(re.compile(r'[/\\\\]legacy_abs_path[/\\\\]', re.IGNORECASE), "Legacy absolute path reference"),
|
|
291
|
+
]
|
|
292
|
+
```
|
|
293
|
+
- **Enterprise Risk**:
|
|
294
|
+
In Python regex syntax, `\b` asserts a word boundary between a word character (`\w`: `[a-zA-Z0-9_]`) and a non-word character (`\W`). Because the forward slash `/` is a non-word character, `\b/Users/` requires that `/` be immediately preceded by an alphanumeric word character. When `/Users/` or `/home/` is preceded by quotation marks (`"`), single quotes (`'`), parentheses (`(`), backticks (`` ` ``), or whitespace, the regex **fails to match**.
|
|
295
|
+
As a result, Gate 2 emitted a false pass while **7 hardcoded absolute developer paths** (`/Users/phungcao/...`) remained in committed repository files:
|
|
296
|
+
1. `docs/PERFORMANCE_AUDIT_REPORT.md:6`: `**Target Repository:** Synapse Enterprise QA Hub (/Users/phungcao/Downloads/Agent Skill)`
|
|
297
|
+
2. `docs/PERFORMANCE_AUDIT_REPORT.md:552`: `... in /Users/phungcao/.gemini/config/plugins/...`
|
|
298
|
+
3. `tests/test_p0_p1_audit_fixes.py:242, 244, 245`: Stack trace mock strings with developer machine paths.
|
|
299
|
+
- **Reproduction Steps**:
|
|
300
|
+
```bash
|
|
301
|
+
python3 -c "
|
|
302
|
+
import re
|
|
303
|
+
from tools.verify_all import FORBIDDEN_PATH_PATTERNS
|
|
304
|
+
test_str = '\"/Users/phungcao/Downloads\"'
|
|
305
|
+
matched = any(p[0].search(test_str) for p in FORBIDDEN_PATH_PATTERNS)
|
|
306
|
+
print('Regex matched quoted path:', matched)
|
|
307
|
+
assert not matched, 'Bypass reproduced!'
|
|
308
|
+
"
|
|
309
|
+
```
|
|
310
|
+
- **Actionable Remediation Proposal**:
|
|
311
|
+
Replace `\b/Users/` and `\b/home/` with a non-word boundary prefix `(?:^|[\s"\'`\(\[\{<])/`:
|
|
312
|
+
```python
|
|
313
|
+
# tools/verify_all.py:93-94
|
|
314
|
+
# BEFORE:
|
|
315
|
+
(re.compile(r'\b/Users/[a-zA-Z0-9_\-]+/', re.IGNORECASE), "macOS/Unix User directory path"),
|
|
316
|
+
(re.compile(r'\b/home/[a-zA-Z0-9_\-]+/', re.IGNORECASE), "Linux User home directory path"),
|
|
317
|
+
|
|
318
|
+
# AFTER:
|
|
319
|
+
(re.compile(r'(?:^|[\s"\'`\(\[\{<])/(?:Users|home)/[a-zA-Z0-9_\-]+/', re.IGNORECASE), "User home directory path (/Users/ or /home/)"),
|
|
320
|
+
```
|
|
321
|
+
Sanitize the 7 leaked instances in `docs/PERFORMANCE_AUDIT_REPORT.md` and `tests/test_p0_p1_audit_fixes.py` with generic placeholders (`/path/to/repo`).
|
|
322
|
+
|
|
323
|
+
---
|
|
324
|
+
|
|
325
|
+
#### Finding SEC-02: Pre-Commit Hook Completely Skips Python Source Files
|
|
326
|
+
- **Severity**: **P1 (High)**
|
|
327
|
+
- **Pillar**: Pillar 5 (Security, Hygiene & Author Attribution Guard)
|
|
328
|
+
- **Exact Location**: `tools/git_pre_commit_guard.py:97-101`
|
|
329
|
+
- **Quoted Source Code**:
|
|
330
|
+
```python
|
|
331
|
+
# tools/git_pre_commit_guard.py:93-101
|
|
332
|
+
def check_forbidden_paths(files):
|
|
333
|
+
violations = []
|
|
334
|
+
for f in files:
|
|
335
|
+
if not os.path.exists(f) or os.path.isdir(f):
|
|
336
|
+
continue
|
|
337
|
+
# Skip this guard script itself and binary files
|
|
338
|
+
if f.endswith((".py", ".db", ".png", ".jpg", ".zip", ".tar", ".gz")):
|
|
339
|
+
continue
|
|
340
|
+
```
|
|
341
|
+
- **Enterprise Risk**:
|
|
342
|
+
Line 99 includes `.py` in the skip list alongside binary files. Consequently, the Git pre-commit hook **never inspects Python source files** for forbidden developer paths, allowing engineers to accidentally commit machine-specific paths in tools, tests, and configuration scripts without pre-commit rejection.
|
|
343
|
+
- **Reproduction Steps**:
|
|
344
|
+
```bash
|
|
345
|
+
python3 -c "
|
|
346
|
+
from tools.git_pre_commit_guard import check_forbidden_paths
|
|
347
|
+
import tempfile, os
|
|
348
|
+
with tempfile.NamedTemporaryFile(suffix='.py', mode='w', delete=False) as tf:
|
|
349
|
+
tf.write('MY_PATH = \"/Users/developer/secret/path\"\n')
|
|
350
|
+
tf_path = tf.name
|
|
351
|
+
try:
|
|
352
|
+
violations = check_forbidden_paths([tf_path])
|
|
353
|
+
print('Violations caught in .py file:', len(violations))
|
|
354
|
+
assert len(violations) == 0, 'Pre-commit hook skipped .py file as predicted'
|
|
355
|
+
finally:
|
|
356
|
+
os.unlink(tf_path)
|
|
357
|
+
"
|
|
358
|
+
```
|
|
359
|
+
- **Actionable Remediation Proposal**:
|
|
360
|
+
In `tools/git_pre_commit_guard.py:99`, remove `.py` from the skip tuple and only skip the guard script itself and true binary extensions:
|
|
361
|
+
```python
|
|
362
|
+
# tools/git_pre_commit_guard.py:98-101
|
|
363
|
+
# BEFORE:
|
|
364
|
+
if f.endswith((".py", ".db", ".png", ".jpg", ".zip", ".tar", ".gz")):
|
|
365
|
+
continue
|
|
366
|
+
|
|
367
|
+
# AFTER:
|
|
368
|
+
if os.path.basename(f) == "git_pre_commit_guard.py" or f.endswith((".db", ".png", ".jpg", ".jpeg", ".zip", ".tar", ".gz", ".ico", ".bin")):
|
|
369
|
+
continue
|
|
370
|
+
```
|
|
371
|
+
|
|
372
|
+
---
|
|
373
|
+
|
|
374
|
+
#### Finding QG-01: Baseline Carte Blanche Whole-File Exception Suppression in Gate 7
|
|
375
|
+
- **Severity**: **P1 (High)**
|
|
376
|
+
- **Pillar**: Pillar 6 (10 Quality Gates Definition of Done)
|
|
377
|
+
- **Exact Location**: `tools/infer_linter.py:253-261` and `tools/infer_linter.py:728-731`
|
|
378
|
+
- **Quoted Source Code**:
|
|
379
|
+
```python
|
|
380
|
+
# tools/infer_linter.py:253-260
|
|
381
|
+
signatures = set()
|
|
382
|
+
for it in items:
|
|
383
|
+
f_path = it.get("file", "").replace("\\", "/")
|
|
384
|
+
d_type = it.get("type", "").upper()
|
|
385
|
+
d_line = it.get("line")
|
|
386
|
+
signatures.add((f_path, d_type))
|
|
387
|
+
if d_line is not None:
|
|
388
|
+
signatures.add((f_path, d_type, d_line))
|
|
389
|
+
|
|
390
|
+
# tools/infer_linter.py:728-731
|
|
391
|
+
# Check baseline suppression
|
|
392
|
+
if baseline_set:
|
|
393
|
+
if (norm_rel, d_type) in baseline_set or (norm_rel, d_type, line_idx) in baseline_set:
|
|
394
|
+
continue
|
|
395
|
+
```
|
|
396
|
+
- **Enterprise Risk**:
|
|
397
|
+
`load_baseline` records two tuples for each baselined item: `(f_path, d_type)` (file-wide) and `(f_path, d_type, d_line)` (line-specific). During verification, if `(norm_rel, d_type) in baseline_set`, the defect is skipped.
|
|
398
|
+
Because 29 core files are listed in `tools/.infer_baseline.json` with `SWALLOWED_EXCEPTION`, **any newly introduced swallowed exception at any line in any of these 29 files is completely suppressed**. An engineer or AI agent could introduce 50 new swallowed exceptions in `tools/core/process_supervisor.py` or `tools/autogen_qa_broker.py`, and Gate 7 will emit a false pass.
|
|
399
|
+
- **Reproduction Steps**:
|
|
400
|
+
```bash
|
|
401
|
+
python3 -c "
|
|
402
|
+
from tools.infer_linter import load_baseline
|
|
403
|
+
baseline = load_baseline('tools/.infer_baseline.json')
|
|
404
|
+
file_wide = ('tools/core/process_supervisor.py', 'SWALLOWED_EXCEPTION') in baseline
|
|
405
|
+
line_fake = ('tools/core/process_supervisor.py', 'SWALLOWED_EXCEPTION', 888888) in baseline or file_wide
|
|
406
|
+
print('Line 888888 suppressed by baseline:', line_fake)
|
|
407
|
+
assert line_fake, 'Bypass confirmed: File-wide carte blanche suppression active'
|
|
408
|
+
"
|
|
409
|
+
```
|
|
410
|
+
- **Actionable Remediation Proposal**:
|
|
411
|
+
Enforce strict line-level or AST-signature matching. Eliminate whole-file tuples `(f_path, d_type)` from `signatures`:
|
|
412
|
+
```python
|
|
413
|
+
# tools/infer_linter.py:253-260
|
|
414
|
+
# BEFORE:
|
|
415
|
+
signatures.add((f_path, d_type))
|
|
416
|
+
if d_line is not None:
|
|
417
|
+
signatures.add((f_path, d_type, d_line))
|
|
418
|
+
|
|
419
|
+
# AFTER:
|
|
420
|
+
if d_line is not None:
|
|
421
|
+
signatures.add((f_path, d_type, int(d_line)))
|
|
422
|
+
|
|
423
|
+
# tools/infer_linter.py:728-731
|
|
424
|
+
# BEFORE:
|
|
425
|
+
if (norm_rel, d_type) in baseline_set or (norm_rel, d_type, line_idx) in baseline_set:
|
|
426
|
+
continue
|
|
427
|
+
|
|
428
|
+
# AFTER:
|
|
429
|
+
if (norm_rel, d_type, line_idx) in baseline_set:
|
|
430
|
+
continue
|
|
431
|
+
```
|
|
432
|
+
|
|
433
|
+
---
|
|
434
|
+
|
|
435
|
+
#### Finding QG-02: Widespread Dummy Exception Bypasses Masking Potential Errors
|
|
436
|
+
- **Severity**: **P1 (High)**
|
|
437
|
+
- **Pillar**: Pillar 6 (10 Quality Gates Definition of Done)
|
|
438
|
+
- **Exact Location**: `tools/verify_all.py:73, 78`, `tools/infer_linter.py:226`, and 27 other files (78 total occurrences)
|
|
439
|
+
- **Quoted Source Code**:
|
|
440
|
+
```python
|
|
441
|
+
# tools/verify_all.py:70-80
|
|
442
|
+
try:
|
|
443
|
+
from tools.rules_router import route_rules
|
|
444
|
+
except ImportError as err:
|
|
445
|
+
_ = err
|
|
446
|
+
try:
|
|
447
|
+
from rules_router import route_rules
|
|
448
|
+
except ImportError as fallback_err:
|
|
449
|
+
_ = fallback_err
|
|
450
|
+
route_rules = None
|
|
451
|
+
|
|
452
|
+
# tools/infer_linter.py:224-227
|
|
453
|
+
except Exception:
|
|
454
|
+
_ = False
|
|
455
|
+
return False
|
|
456
|
+
```
|
|
457
|
+
- **Enterprise Risk**:
|
|
458
|
+
A global AST scan across `tools/` and `tests/` revealed **78 occurrences** of dummy assignments (`_ = err`, `_ = False`, `const _ = e`). This pattern was originally introduced to satisfy naive AST linters that flag empty `except:` blocks.
|
|
459
|
+
However, assigning to `_` directly violates AGENTS.md § 2.3 ("No dummy bypasses `_ = err`, `_ = False`, empty catch") and AGENTS.md § 3.1 ("Truth Anchor: Zero Fake Passes"). It silently swallows critical system exceptions (e.g. underlying file permission errors, broken import chains, corrupted database states), preventing developers from diagnosing root causes.
|
|
460
|
+
- **Reproduction Steps**:
|
|
461
|
+
```bash
|
|
462
|
+
python3 -c "
|
|
463
|
+
import subprocess
|
|
464
|
+
res = subprocess.run(['git', 'grep', '-n', '-E', r'_\s*=\s*(err|exc|e|fallback_err|False)\b', 'tools/'], capture_output=True, text=True)
|
|
465
|
+
lines = res.stdout.strip().split('\n')
|
|
466
|
+
print(f'Total dummy bypass occurrences in tools/: {len(lines)}')
|
|
467
|
+
assert len(lines) > 50, 'Dummy bypass occurrences verified'
|
|
468
|
+
"
|
|
469
|
+
```
|
|
470
|
+
- **Actionable Remediation Proposal**:
|
|
471
|
+
Replace dummy assignments with explicit debug logging or structured diagnostics:
|
|
472
|
+
```python
|
|
473
|
+
# tools/verify_all.py:70-80
|
|
474
|
+
# BEFORE:
|
|
475
|
+
except ImportError as err:
|
|
476
|
+
_ = err
|
|
477
|
+
try:
|
|
478
|
+
from rules_router import route_rules
|
|
479
|
+
except ImportError as fallback_err:
|
|
480
|
+
_ = fallback_err
|
|
481
|
+
route_rules = None
|
|
482
|
+
|
|
483
|
+
# AFTER:
|
|
484
|
+
except ImportError:
|
|
485
|
+
try:
|
|
486
|
+
from rules_router import route_rules
|
|
487
|
+
except ImportError as fallback_err:
|
|
488
|
+
import logging
|
|
489
|
+
logging.getLogger("verify_all").debug("Failed to import rules_router: %s", fallback_err)
|
|
490
|
+
route_rules = None
|
|
491
|
+
```
|
|
492
|
+
|
|
493
|
+
---
|
|
494
|
+
|
|
495
|
+
#### Finding F-P2-01: Sub-3000ms Static Sleeps Pass Gate 7 Undetected
|
|
496
|
+
- **Severity**: **P1 (High)**
|
|
497
|
+
- **Pillar**: Pillar 2 (Clean POM Flat Architecture & Test Structure)
|
|
498
|
+
- **Exact Location**: `tools/infer_linter.py:447-449` and `tools/infer_linter.py:475-477`
|
|
499
|
+
- **Quoted Source Code**:
|
|
500
|
+
```python
|
|
501
|
+
# tools/infer_linter.py:447-449
|
|
502
|
+
is_excessive = delay_ms >= 3000
|
|
503
|
+
d_type = "EXCESSIVE_HARD_DELAY" if is_excessive else "HARD_DELAY_ANTI_PATTERN"
|
|
504
|
+
sev = "CRITICAL" if (is_excessive and role == "TEST_SCRIPT") else "WARNING"
|
|
505
|
+
|
|
506
|
+
# tools/infer_linter.py:942-945
|
|
507
|
+
if strict:
|
|
508
|
+
fake_pass_count = sum(1 for r in results for d in r["defects"] if d["type"] in ("FAKE_PASS_NO_ASSERTION", "NO_ASSERTION_IN_TEST", "DANGLING_ASSERTION_NO_MATCHER"))
|
|
509
|
+
return 0 if (critical_count == 0 and fake_pass_count == 0) else 1
|
|
510
|
+
return 0 if critical_count == 0 else 1
|
|
511
|
+
```
|
|
512
|
+
- **Enterprise Risk**:
|
|
513
|
+
`infer_linter.py` assigns `severity="WARNING"` to static delays under 3000ms. In both standard and `--strict` mode, `run_infer_linter` only returns an exit code of 1 if `critical_count > 0` or fake-pass assertions are found.
|
|
514
|
+
Consequently, test scripts containing `time.sleep(1)`, `Thread.sleep(1500)`, or `page.waitForTimeout(2000)` **exit with code 0 (PASS)**. This directly contradicts AGENTS.md § 1.4: *"Static sleeps (Thread.sleep, time.sleep) are strictly prohibited to prevent pipeline flakiness."* It allows flaky timing debt to proliferate in CI pipelines.
|
|
515
|
+
- **Reproduction Steps**:
|
|
516
|
+
```bash
|
|
517
|
+
python3 -c "
|
|
518
|
+
import tempfile, os, subprocess, sys
|
|
519
|
+
with tempfile.TemporaryDirectory() as td:
|
|
520
|
+
p = os.path.join(td, 'test_sleep.py')
|
|
521
|
+
with open(p, 'w') as f:
|
|
522
|
+
f.write('import time\ndef test_short_sleep():\n time.sleep(1)\n assert True\n')
|
|
523
|
+
res = subprocess.run([sys.executable, 'tools/infer_linter.py', '--strict', '--dir', td], capture_output=True, text=True)
|
|
524
|
+
print('Exit code for time.sleep(1):', res.returncode)
|
|
525
|
+
assert res.returncode == 0, 'Bypass reproduced: 1s sleep passed Gate 7'
|
|
526
|
+
"
|
|
527
|
+
```
|
|
528
|
+
- **Actionable Remediation Proposal**:
|
|
529
|
+
In `tools/infer_linter.py`, when scanning files with `role == "TEST_SCRIPT"`, classify **any static sleep** as `CRITICAL`:
|
|
530
|
+
```python
|
|
531
|
+
# tools/infer_linter.py:447-449
|
|
532
|
+
# BEFORE:
|
|
533
|
+
is_excessive = delay_ms >= 3000
|
|
534
|
+
d_type = "EXCESSIVE_HARD_DELAY" if is_excessive else "HARD_DELAY_ANTI_PATTERN"
|
|
535
|
+
sev = "CRITICAL" if (is_excessive and role == "TEST_SCRIPT") else "WARNING"
|
|
536
|
+
|
|
537
|
+
# AFTER:
|
|
538
|
+
is_excessive = delay_ms >= 3000
|
|
539
|
+
d_type = "EXCESSIVE_HARD_DELAY" if is_excessive else "HARD_DELAY_ANTI_PATTERN"
|
|
540
|
+
sev = "CRITICAL" if role == "TEST_SCRIPT" else ("CRITICAL" if is_excessive else "WARNING")
|
|
541
|
+
```
|
|
542
|
+
|
|
543
|
+
---
|
|
544
|
+
|
|
545
|
+
#### Finding F-P2-02: Gate 3 Clean POM Linter Blind to Playwright Semantic Locators
|
|
546
|
+
- **Severity**: **P1 (High)**
|
|
547
|
+
- **Pillar**: Pillar 2 (Clean POM Flat Architecture & Test Structure)
|
|
548
|
+
- **Exact Location**: `tools/verify_all.py:373-401`
|
|
549
|
+
- **Quoted Source Code**:
|
|
550
|
+
```python
|
|
551
|
+
# tools/verify_all.py:373-379
|
|
552
|
+
LOCATOR_CALL_PATTERN = re.compile(
|
|
553
|
+
r'(?:\.(?:click|dblclick|fill|type|press|check|locator|waitForSelector|hover|selectOption)|cy\.(?:get|find)|driver\.find[A-Za-z]*|WebUI\.[a-zA-Z]+)\s*\(\s*[\x22\x27`](?P<sel>[^\x22\x27`]+)[\x22\x27`]'
|
|
554
|
+
)
|
|
555
|
+
RAW_XPATH_PATTERN = re.compile(r'[\x22\x27`](?:xpath=|\/{2}[a-zA-Z*@\[])[^\x22\x27`]*[\x22\x27`]')
|
|
556
|
+
```
|
|
557
|
+
- **Enterprise Risk**:
|
|
558
|
+
`gate_3_clean_pom` validates that test scripts do not call raw DOM locators. However, `LOCATOR_CALL_PATTERN` only catches traditional method calls whose arguments look like CSS/XPath selectors (starting with `.`, `#`, `//`, `xpath=`, or `[`).
|
|
559
|
+
Modern Playwright tests rely on semantic locators: `page.getByRole(...)`, `page.getByLabel(...)`, `page.getByTestId(...)`, `page.getByText(...)`, and `page.getByPlaceholder(...)`. When these calls are placed directly in test scripts outside of Page Classes, **Gate 3 ignores them completely** and reports 100% Clean POM compliance. This allows architectural erosion where DOM knowledge leaks into test scripts.
|
|
560
|
+
- **Reproduction Steps**:
|
|
561
|
+
```bash
|
|
562
|
+
python3 -c "
|
|
563
|
+
import tempfile, os, sys
|
|
564
|
+
sys.path.insert(0, 'tools')
|
|
565
|
+
from verify_all import gate_3_clean_pom
|
|
566
|
+
with tempfile.TemporaryDirectory() as td:
|
|
567
|
+
tests_dir = os.path.join(td, 'tests')
|
|
568
|
+
os.makedirs(tests_dir)
|
|
569
|
+
with open(os.path.join(tests_dir, 'auth.spec.ts'), 'w') as f:
|
|
570
|
+
f.write('test(\"bad pom\", async ({page}) => { await page.getByRole(\"button\", {name: \"Login\"}).click(); });\n')
|
|
571
|
+
passed = gate_3_clean_pom(td, mode='TARGET')
|
|
572
|
+
print('Gate 3 passed unencapsulated getByRole:', passed)
|
|
573
|
+
assert passed is True, 'Blindspot confirmed: Semantic locator in test script bypassed Gate 3'
|
|
574
|
+
"
|
|
575
|
+
```
|
|
576
|
+
- **Actionable Remediation Proposal**:
|
|
577
|
+
Add a dedicated semantic locator inspection regex to `gate_3_clean_pom`:
|
|
578
|
+
```python
|
|
579
|
+
# tools/verify_all.py:380
|
|
580
|
+
SEMANTIC_LOCATOR_PATTERN = re.compile(
|
|
581
|
+
r'(?:page|frame)\.(?:getByRole|getByLabel|getByPlaceholder|getByText|getByTestId|getByAltText|getByTitle)\s*\('
|
|
582
|
+
)
|
|
583
|
+
|
|
584
|
+
# Inside the file inspection loop in gate_3_clean_pom:
|
|
585
|
+
if SEMANTIC_LOCATOR_PATTERN.search(stripped) and not is_page_class_path(norm_path):
|
|
586
|
+
violations.append(f"{rel_path}:{idx}: Raw semantic locator call '{stripped.strip()}' in test script; encapsulate into a Page Class method.")
|
|
587
|
+
```
|
|
588
|
+
|
|
589
|
+
---
|
|
590
|
+
|
|
591
|
+
#### Finding SEC-03: Incomplete Author Attribution & SPDX Header Coverage
|
|
592
|
+
- **Severity**: **P2 (Medium)**
|
|
593
|
+
- **Pillar**: Pillar 5 (Security, Hygiene & Author Attribution Guard)
|
|
594
|
+
- **Exact Location**: Repository source files (`tools/`, `tests/`) and `tools/author_integrity.py:15-30`
|
|
595
|
+
- **Quoted Source Code**:
|
|
596
|
+
```python
|
|
597
|
+
# tools/author_integrity.py:24-30
|
|
598
|
+
REQUIRED_METADATA_FILES = [
|
|
599
|
+
"LICENSE",
|
|
600
|
+
"package.json",
|
|
601
|
+
"README.md",
|
|
602
|
+
"pyproject.toml",
|
|
603
|
+
]
|
|
604
|
+
```
|
|
605
|
+
- **Enterprise Risk**:
|
|
606
|
+
`AGENTS.md` Rule 0 explicitly states: *"Legally, the MIT License mandates that original copyright notices and permission notices MUST be preserved in all copies or substantial portions of the Software (`LICENSE` and code headers)."*
|
|
607
|
+
However, an AST scan across 282 repository source files revealed:
|
|
608
|
+
- Files containing author attribution (`Thế Vị` / `SPDX-FileCopyrightText`): **150 files (53.2%)**
|
|
609
|
+
- Files missing copyright headers: **132 files (46.8%)** (including core utilities `tools/infer_linter.py`, `tools/setup_project.py`, `tools/repo_map.py`, `tools/wcag_audit.py`).
|
|
610
|
+
Because `tools/author_integrity.py` only validates 4 root-level files (`LICENSE`, `package.json`, `README.md`, `pyproject.toml`), third-party forks can strip headers from internal source files without triggering integrity failures.
|
|
611
|
+
- **Reproduction Steps**:
|
|
612
|
+
```bash
|
|
613
|
+
python3 -c "
|
|
614
|
+
import os
|
|
615
|
+
srcs = [os.path.join(r, f) for r, _, fs in os.walk('tools') for f in fs if f.endswith('.py')]
|
|
616
|
+
missing = [f for f in srcs if 'Thế Vị' not in open(f, encoding='utf-8', errors='ignore').read()]
|
|
617
|
+
print(f'Missing headers in tools/: {len(missing)} / {len(srcs)} files')
|
|
618
|
+
assert len(missing) > 0
|
|
619
|
+
"
|
|
620
|
+
```
|
|
621
|
+
- **Actionable Remediation Proposal**:
|
|
622
|
+
Add an automated header verification function in `tools/author_integrity.py` that checks for the standard SPDX header on all `.py`, `.ts`, and `.groovy` files:
|
|
623
|
+
```python
|
|
624
|
+
# SPDX-FileCopyrightText: © 2026 Thế Vị (@TeeVee98) <https://github.com/TeeVee98/Synapse-Hub>
|
|
625
|
+
# SPDX-License-Identifier: MIT
|
|
626
|
+
```
|
|
627
|
+
|
|
628
|
+
---
|
|
629
|
+
|
|
630
|
+
#### Finding SEC-04: Pre-Commit Hook Lacks Secret Scanning and Author Integrity Pre-Flight
|
|
631
|
+
- **Severity**: **P2 (Medium)**
|
|
632
|
+
- **Pillar**: Pillar 5 (Security, Hygiene & Author Attribution Guard)
|
|
633
|
+
- **Exact Location**: `tools/git_pre_commit_guard.py:120-145`
|
|
634
|
+
- **Quoted Source Code**:
|
|
635
|
+
```python
|
|
636
|
+
# tools/git_pre_commit_guard.py:120-145
|
|
637
|
+
def main():
|
|
638
|
+
files = get_staged_files()
|
|
639
|
+
# Only runs check_forbidden_paths and temporary file blacklist checks
|
|
640
|
+
```
|
|
641
|
+
- **Enterprise Risk**:
|
|
642
|
+
The Git pre-commit hook only verifies forbidden path patterns and temporary file names. It completely omits Gate 6 (Credential Security / Secret Scanning) and Author Attribution verification. Developers can stage and commit AWS secret keys, GitHub personal access tokens, or private RSA keys to local Git history before running `verify_all.py`.
|
|
643
|
+
- **Reproduction Steps**:
|
|
644
|
+
Inspect `tools/git_pre_commit_guard.py`. Note the total absence of `FORBIDDEN_CREDENTIAL_PATTERNS` or `author_integrity` invocations.
|
|
645
|
+
- **Actionable Remediation Proposal**:
|
|
646
|
+
Import and invoke `gate_6_credential_security` and `assert_integrity_or_raise` from inside `tools/git_pre_commit_guard.py:main()`.
|
|
647
|
+
|
|
648
|
+
---
|
|
649
|
+
|
|
650
|
+
#### Finding QG-03: `safe_print` Stream Redirection Default Argument Leaks to Stdout During JSON Export
|
|
651
|
+
- **Severity**: **P2 (Medium)**
|
|
652
|
+
- **Pillar**: Pillar 6 (10 Quality Gates Definition of Done)
|
|
653
|
+
- **Exact Location**: `tools/verify_all.py:58, 851, 877, 1259`
|
|
654
|
+
- **Quoted Source Code**:
|
|
655
|
+
```python
|
|
656
|
+
# tools/verify_all.py:58
|
|
657
|
+
def safe_print(text="", stream=sys.stdout):
|
|
658
|
+
try:
|
|
659
|
+
if text:
|
|
660
|
+
stream.write(f"{text}\n")
|
|
661
|
+
...
|
|
662
|
+
# tools/verify_all.py:1259
|
|
663
|
+
if output_format == "json":
|
|
664
|
+
sys.stdout = io.StringIO()
|
|
665
|
+
```
|
|
666
|
+
- **Enterprise Risk**:
|
|
667
|
+
In Python, default arguments are evaluated once at module definition time. In `safe_print`, `stream=sys.stdout` binds to the original terminal `sys.stdout`. Later, when `run_all_gates` redirects `sys.stdout = io.StringIO()` for JSON output, calls in Gate 7 (lines 851, 877) that call `safe_print(...)` continue writing directly to the underlying terminal stream.
|
|
668
|
+
Running `python3 tools/verify_all.py --format json` outputs raw text banners mixed with JSON:
|
|
669
|
+
```text
|
|
670
|
+
[GATE 7] Executing Hub Regression Test Suite (pytest tests -q)...
|
|
671
|
+
[PASS] 100% Hub regression test suite passed.
|
|
672
|
+
{ "timestamp": "...", "gates": ... }
|
|
673
|
+
```
|
|
674
|
+
Automated CI/CD pipelines parsing stdout with `jq` or `json.loads` crash with `JSONDecodeError`.
|
|
675
|
+
- **Reproduction Steps**:
|
|
676
|
+
```bash
|
|
677
|
+
python3 tools/verify_all.py --format json 2>&1 | python3 -c "import sys, json; json.loads(sys.stdin.read())"
|
|
678
|
+
# Crashes with: json.decoder.JSONDecodeError: Expecting value: line 1 column 2 (char 1)
|
|
679
|
+
```
|
|
680
|
+
- **Actionable Remediation Proposal**:
|
|
681
|
+
Change the default parameter to `stream=None` and resolve `sys.stdout` dynamically:
|
|
682
|
+
```python
|
|
683
|
+
# tools/verify_all.py:58
|
|
684
|
+
# BEFORE:
|
|
685
|
+
def safe_print(text="", stream=sys.stdout):
|
|
686
|
+
|
|
687
|
+
# AFTER:
|
|
688
|
+
def safe_print(text="", stream=None):
|
|
689
|
+
target_stream = sys.stdout if stream is None else stream
|
|
690
|
+
try:
|
|
691
|
+
if text:
|
|
692
|
+
target_stream.write(f"{text}\n")
|
|
693
|
+
else:
|
|
694
|
+
target_stream.write("\n")
|
|
695
|
+
target_stream.flush()
|
|
696
|
+
except UnicodeEncodeError:
|
|
697
|
+
target_stream.write(text.encode("ascii", "backslashreplace").decode("ascii") + "\n")
|
|
698
|
+
```
|
|
699
|
+
|
|
700
|
+
---
|
|
701
|
+
|
|
702
|
+
#### Finding F-P2-03: Selenium Reference Suite Lacks Automated Teardown Cleanup
|
|
703
|
+
- **Severity**: **P2 (Medium)**
|
|
704
|
+
- **Pillar**: Pillar 2 (Clean POM Flat Architecture & Test Structure)
|
|
705
|
+
- **Exact Location**: `examples/selenium-suite/tests/test_order.py:4-10`
|
|
706
|
+
- **Quoted Source Code**:
|
|
707
|
+
```python
|
|
708
|
+
# examples/selenium-suite/tests/test_order.py:4-10
|
|
709
|
+
def test_order_creation_and_receipt():
|
|
710
|
+
"""Validates idempotent order checkout flow."""
|
|
711
|
+
order_page = OrderPage()
|
|
712
|
+
order_ref = order_page.create_order(sku="PROD_TEST_99", quantity=2)
|
|
713
|
+
|
|
714
|
+
assert order_ref.startswith("ORD_")
|
|
715
|
+
assert order_page.verify_order_placed(order_ref) is True
|
|
716
|
+
```
|
|
717
|
+
- **Enterprise Risk**:
|
|
718
|
+
`examples/playwright-crm-suite` and `examples/pytest-api-suite` implement teardown registries that delete test entities after execution. In contrast, `examples/selenium-suite/tests/test_order.py` creates an order without any teardown fixture or deletion. This provides an incomplete reference model for engineers copying from `examples/selenium-suite`, violating AGENTS.md § 1.3: *"Every creation test case must be paired with automated Teardown cleanup."*
|
|
719
|
+
- **Reproduction Steps**:
|
|
720
|
+
Inspect `examples/selenium-suite/tests/test_order.py:4-10`. Note the absence of `pytest.fixture` teardown, `test.afterEach`, or `finally:` block.
|
|
721
|
+
- **Actionable Remediation Proposal**:
|
|
722
|
+
Add an automated teardown fixture in `examples/selenium-suite/conftest.py` that registers created order references and deletes or cancels them in teardown.
|
|
723
|
+
|
|
724
|
+
---
|
|
725
|
+
|
|
726
|
+
#### Finding F-P3-01: Cognitive-Only External Plugin Suppression
|
|
727
|
+
- **Severity**: **P2 (Medium)**
|
|
728
|
+
- **Pillar**: Pillar 3 (Multi-Agent Swarm & Finite State Machine)
|
|
729
|
+
- **Exact Location**: `.agents/rules/skill-precedence.md:5-20`
|
|
730
|
+
- **Enterprise Risk**:
|
|
731
|
+
Skill precedence (`Platform Guardrails > Core Rules > Local Skills > External Plugins`) is implemented as in-context behavioral instructions. If a host harness (e.g. Claude Code or Gemini CLI) mounts third-party MCP plugins or external tool definitions, suppression relies on the LLM's cognitive compliance rather than OS-level process sandboxing or tool schema filtering. A non-compliant model could invoke external plugins when local skills are available.
|
|
732
|
+
- **Reproduction Steps**:
|
|
733
|
+
Inspect `.agents/rules/skill-precedence.md`. Note that rules specify behavioral guidance rather than programmatic runtime MCP tool removal.
|
|
734
|
+
- **Actionable Remediation Proposal**:
|
|
735
|
+
Extend `tools/fsm_middleware.py` or MCP dispatch proxies to dynamically filter third-party tool manifests out of the LLM tool declaration list when project-local skill equivalents exist.
|
|
736
|
+
|
|
737
|
+
---
|
|
738
|
+
|
|
739
|
+
#### Finding F-P7-01: Unhandled `FileNotFoundError` Crash in `tools/project_state.py` on Subcommand Argument
|
|
740
|
+
- **Severity**: **P2 (Medium)**
|
|
741
|
+
- **Pillar**: Pillar 7 (Developer Experience & Zero-Command Invariant)
|
|
742
|
+
- **Exact Location**: `tools/project_state.py:169-170`
|
|
743
|
+
- **Quoted Source Code**:
|
|
744
|
+
```python
|
|
745
|
+
# tools/project_state.py:169-170
|
|
746
|
+
target = sys.argv[1] if len(sys.argv) > 1 else os.getcwd()
|
|
747
|
+
sf = init_state(target)
|
|
748
|
+
```
|
|
749
|
+
- **Enterprise Risk**:
|
|
750
|
+
When an engineer or automated agent executes `python3 tools/project_state.py status` or `python3 tools/project_state.py read`, line 169 treats the string `"status"` as a directory path and passes it to `init_state("status")`. Because the folder `status/` does not exist, `open("status/PROJECT_STATE.md", "w")` crashes with an unhandled `FileNotFoundError: [Errno 2] No such file or directory: 'status/PROJECT_STATE.md'`.
|
|
751
|
+
- **Reproduction Steps**:
|
|
752
|
+
```bash
|
|
753
|
+
python3 tools/project_state.py status
|
|
754
|
+
# Crashes with: FileNotFoundError: [Errno 2] No such file or directory: 'status/PROJECT_STATE.md'
|
|
755
|
+
```
|
|
756
|
+
- **Actionable Remediation Proposal**:
|
|
757
|
+
Parse subcommands in `tools/project_state.py:165-175`:
|
|
758
|
+
```python
|
|
759
|
+
# tools/project_state.py:169-175
|
|
760
|
+
# BEFORE:
|
|
761
|
+
target = sys.argv[1] if len(sys.argv) > 1 else os.getcwd()
|
|
762
|
+
sf = init_state(target)
|
|
763
|
+
|
|
764
|
+
# AFTER:
|
|
765
|
+
arg = sys.argv[1] if len(sys.argv) > 1 else "read"
|
|
766
|
+
if arg in ("status", "read", "view"):
|
|
767
|
+
print(read_state(os.getcwd()))
|
|
768
|
+
elif os.path.isdir(arg):
|
|
769
|
+
sf = init_state(arg)
|
|
770
|
+
print(f"State initialized at: {sf}")
|
|
771
|
+
else:
|
|
772
|
+
sf = init_state(os.getcwd())
|
|
773
|
+
print(f"State initialized at: {sf}")
|
|
774
|
+
```
|
|
775
|
+
|
|
776
|
+
---
|
|
777
|
+
|
|
778
|
+
#### Finding ROUTER-01: Stack Marker Directory Recursion Truncated at Depth 2
|
|
779
|
+
- **Severity**: **P2 (Medium)**
|
|
780
|
+
- **Pillar**: Pillar 1 (Token Engineering & Context Optimization)
|
|
781
|
+
- **Exact Location**: `tools/rules_router.py:306-314`
|
|
782
|
+
- **Quoted Source Code**:
|
|
783
|
+
```python
|
|
784
|
+
# tools/rules_router.py:306-314
|
|
785
|
+
def _scan_directory_markers(base_dir: str, depth: int = 0) -> Set[str]:
|
|
786
|
+
if depth >= 2:
|
|
787
|
+
return set()
|
|
788
|
+
found = set()
|
|
789
|
+
# Iterates directory entries...
|
|
790
|
+
```
|
|
791
|
+
- **Enterprise Risk**:
|
|
792
|
+
To prevent unbounded traversal on large codebases, `_scan_directory_markers` truncates directory recursion at `depth >= 2`. In deeply nested enterprise monorepos (e.g. `services/core/nested/deep/package.json`), stack markers located deeper than depth 2 from the workspace root or top-level subpackages (`packages/`, `apps/`, `services/`) will not be detected. The router falls back to unclassified workspace status.
|
|
793
|
+
- **Reproduction Steps**:
|
|
794
|
+
```bash
|
|
795
|
+
python3 -c "
|
|
796
|
+
import tempfile, os
|
|
797
|
+
from tools.rules_router import detect_stacks
|
|
798
|
+
with tempfile.TemporaryDirectory() as td:
|
|
799
|
+
deep_dir = os.path.join(td, 'level1', 'level2', 'level3')
|
|
800
|
+
os.makedirs(deep_dir)
|
|
801
|
+
with open(os.path.join(deep_dir, 'playwright.config.ts'), 'w') as f:
|
|
802
|
+
f.write('export default {};\n')
|
|
803
|
+
stacks = detect_stacks(td)
|
|
804
|
+
print('Detected stacks in deep monorepo:', stacks)
|
|
805
|
+
assert stacks == ['unknown'], 'Depth-2 truncation confirmed'
|
|
806
|
+
"
|
|
807
|
+
```
|
|
808
|
+
- **Actionable Remediation Proposal**:
|
|
809
|
+
Expose a configurable `--depth <int>` CLI parameter (default: 2, max: 4) or inspect root workspace manifests (`pnpm-workspace.yaml`, `lerna.json`, `pom.xml`, `settings.gradle`) to resolve subproject directories dynamically.
|
|
810
|
+
|
|
811
|
+
---
|
|
812
|
+
|
|
813
|
+
## 4. ENTERPRISE PRODUCTION READINESS VERDICT
|
|
814
|
+
|
|
815
|
+
### 4.1 Production Readiness Verdict
|
|
816
|
+
|
|
817
|
+
$$\mathbf{\text{VERDICT: CONDITIONAL ENTERPRISE PRODUCTION READY}}$$
|
|
818
|
+
|
|
819
|
+
Synapse Hub is **authorized for enterprise staging and pre-production deployments**. Its core architecture — including the 5-Stage Finite State Machine, Dual-Tier Token Hydration, Noah Shinn Reflexion Self-Healing, Reversible Sandbox WAL, and Polyglot Version Synchronizer — is mature, robust, and outperforms typical enterprise automation suites.
|
|
820
|
+
|
|
821
|
+
However, **mission-critical enterprise production deployment requires resolving the 6 P1 findings** to eliminate linter blindspots and ensure strict gate integrity.
|
|
822
|
+
|
|
823
|
+
---
|
|
824
|
+
|
|
825
|
+
### 4.2 Pre-Deployment Gate Conditions (The 6 P1 Blockers)
|
|
826
|
+
|
|
827
|
+
Prior to promoting Synapse Hub to unassisted production CI/CD pipelines, enterprise platform teams must enforce the following pre-deployment gate conditions:
|
|
828
|
+
|
|
829
|
+
1. **Gate 2 Path Portability Regex Fix (SEC-01)**:
|
|
830
|
+
Update `FORBIDDEN_PATH_PATTERNS` in `tools/verify_all.py` to use `(?:^|[\s"\'`\(\[\{<])/` and sanitize the 7 leaked developer path occurrences in documentation and test mock strings.
|
|
831
|
+
2. **Pre-Commit Hook Python Scanning (SEC-02)**:
|
|
832
|
+
Remove `.py` from the skip list in `tools/git_pre_commit_guard.py:99` so that Python source files are verified for forbidden paths before staging.
|
|
833
|
+
3. **Gate 7 Baseline Line-Specific Enforcement (QG-01)**:
|
|
834
|
+
Refactor `tools/infer_linter.py:258, 730` to require exact line matches `(f_path, d_type, d_line)` for baseline suppression, eliminating file-wide carte blanche exception suppression.
|
|
835
|
+
4. **Remediate Internal Dummy Bypasses (QG-02)**:
|
|
836
|
+
Refactor dummy exception assignments (`_ = err`, `_ = False`) across `tools/` and `tests/` to use explicit logger calls or typed exception handlers.
|
|
837
|
+
5. **Enforce Static Sleep Prohibition in Test Scripts (F-P2-01)**:
|
|
838
|
+
Modify `tools/infer_linter.py:447-449` so that any static sleep within test scripts is classified as `CRITICAL`, failing Gate 7 when static delays (< 3000ms) are present.
|
|
839
|
+
6. **Expand Gate 3 Clean POM for Playwright Locators (F-P2-02)**:
|
|
840
|
+
Add `SEMANTIC_LOCATOR_PATTERN` to `tools/verify_all.py:373` to catch unencapsulated `page.getByRole(...)`, `page.getByLabel(...)`, and `page.getByTestId(...)` in test scripts.
|
|
841
|
+
|
|
842
|
+
---
|
|
843
|
+
|
|
844
|
+
### 4.3 3-Phase Enterprise Remediation Roadmap
|
|
845
|
+
|
|
846
|
+
```
|
|
847
|
+
Phase 1: Immediate Security & Gate Hotfixes (Sprint 1 — Days 1 to 3)
|
|
848
|
+
├── 1.1 Patch Gate 2 regex in tools/verify_all.py (SEC-01)
|
|
849
|
+
├── 1.2 Remove .py skip in tools/git_pre_commit_guard.py (SEC-02)
|
|
850
|
+
├── 1.3 Update infer_linter.py to classify any test sleep as CRITICAL (F-P2-01)
|
|
851
|
+
├── 1.4 Add Playwright semantic locator regex to Gate 3 (F-P2-02)
|
|
852
|
+
└── 1.5 Fix safe_print default parameter for valid JSON output (QG-03)
|
|
853
|
+
|
|
854
|
+
Phase 2: Static Defect & Teardown Cleanup (Sprint 2 — Days 4 to 7)
|
|
855
|
+
├── 2.1 Refactor infer_linter.py baseline to line-specific matching (QG-01)
|
|
856
|
+
├── 2.2 Clean up 78 dummy bypasses (_ = err) across tools/ (QG-02)
|
|
857
|
+
├── 2.3 Add automated teardown fixture to examples/selenium-suite (F-P2-03)
|
|
858
|
+
└── 2.4 Add SPDX author headers to the 132 missing source files (SEC-03)
|
|
859
|
+
|
|
860
|
+
Phase 3: Ergonomics & Monorepo Scaling (Sprint 3 — Days 8 to 14)
|
|
861
|
+
├── 3.1 Fix subcommand argument parsing in tools/project_state.py (F-P7-01)
|
|
862
|
+
├── 3.2 Add configurable scan depth (--depth) to tools/rules_router.py (ROUTER-01)
|
|
863
|
+
├── 3.3 Integrate secret scanning and author checks into pre-commit guard (SEC-04)
|
|
864
|
+
└── 3.4 Implement runtime MCP tool schema filtering for external plugins (F-P3-01)
|
|
865
|
+
```
|
|
866
|
+
|
|
867
|
+
---
|
|
868
|
+
|
|
869
|
+
## 5. APPENDIX: REPRODUCTION COMMAND SUITE & VERIFICATION GUIDE
|
|
870
|
+
|
|
871
|
+
This appendix provides a turn-key suite of terminal commands to verify all findings and test suites reported in this audit.
|
|
872
|
+
|
|
873
|
+
---
|
|
874
|
+
|
|
875
|
+
### 5.1 Verification Commands for Passing Pillars & Core Suites
|
|
876
|
+
|
|
877
|
+
Execute these commands in the repository root to verify the passing behavior of core components:
|
|
878
|
+
|
|
879
|
+
```bash
|
|
880
|
+
# 1. Verify 10 Quality Gates DoD Runner (Pillar 6)
|
|
881
|
+
python3 tools/verify_all.py
|
|
882
|
+
|
|
883
|
+
# 2. Verify Pillar 1 & 4 Test Suite (71 Tests: Benchmark, Token Estimator, Rules Router, Self-Healing, WAL, Mutator)
|
|
884
|
+
python3 -m pytest tests/test_benchmark_context.py tests/test_token_estimator.py tests/test_rules_router.py tests/test_distill_log.py tests/test_auto_heal_reflexion.py tests/test_reversible_sandbox.py tests/test_reversible_sandbox_wal.py tests/test_reversible_sandbox_fault_injection.py tests/test_micro_mutator.py -q
|
|
885
|
+
|
|
886
|
+
# 3. Verify Pillar 2, 3 & 7 Test Suite (49 Tests: FSM Enforcer, Middleware, POM Linter, Bump Version, Hub Publisher)
|
|
887
|
+
python3 -m pytest tests/test_fsm_enforcer.py tests/test_fsm_middleware.py tests/test_pom_adversarial_linter.py tests/test_bump_version.py tests/test_hub_publisher.py -q
|
|
888
|
+
|
|
889
|
+
# 4. Verify 100% Polyglot Version Synchronization Across 20 Targets (Pillar 7)
|
|
890
|
+
python3 tools/check_version_sync.py
|
|
891
|
+
|
|
892
|
+
# 5. Verify Subagent Trigger & Skill Collision Guard (Pillar 3)
|
|
893
|
+
python3 tools/skill_collision_guard.py --dir .agents/skills --threshold 0.75 --scan-rules --rules-dir .agents/rules
|
|
894
|
+
|
|
895
|
+
# 6. Verify 100% UTF-8 No BOM Encoding Integrity across all 660 candidate files (Pillar 5)
|
|
896
|
+
python3 -c "
|
|
897
|
+
import os
|
|
898
|
+
exts = ('.groovy', '.java', '.ts', '.py', '.json', '.md', '.tc', '.glbl', '.yaml', '.yml', '.xml')
|
|
899
|
+
boms = [os.path.join(r, f) for r, _, fs in os.walk('.') if not any(x in r for x in ['.git', 'node_modules', 'teamwork']) for f in fs if f.endswith(exts) and open(os.path.join(r, f), 'rb').read(3) == b'\xef\xbb\xbf']
|
|
900
|
+
print('BOM Violations:', len(boms))
|
|
901
|
+
assert len(boms) == 0
|
|
902
|
+
"
|
|
903
|
+
```
|
|
904
|
+
|
|
905
|
+
---
|
|
906
|
+
|
|
907
|
+
### 5.2 Standalone Finding Reproduction Commands
|
|
908
|
+
|
|
909
|
+
#### A. Reproduce SEC-01 (Gate 2 Regex Word-Boundary Flaw)
|
|
910
|
+
```bash
|
|
911
|
+
python3 -c "
|
|
912
|
+
from tools.verify_all import FORBIDDEN_PATH_PATTERNS
|
|
913
|
+
path = '\"/Users/phungcao/Downloads\"'
|
|
914
|
+
matched = any(p[0].search(path) for p in FORBIDDEN_PATH_PATTERNS)
|
|
915
|
+
print('SEC-01: Quoted path detected by Gate 2 regex:', matched)
|
|
916
|
+
assert not matched, 'Bypass verified: Quoted absolute path slipped through'
|
|
917
|
+
"
|
|
918
|
+
```
|
|
919
|
+
|
|
920
|
+
#### B. Reproduce SEC-02 (Pre-Commit Hook Skipping Python Files)
|
|
921
|
+
```bash
|
|
922
|
+
python3 -c "
|
|
923
|
+
from tools.git_pre_commit_guard import check_forbidden_paths
|
|
924
|
+
import tempfile, os
|
|
925
|
+
with tempfile.NamedTemporaryFile(suffix='.py', mode='w', delete=False) as f:
|
|
926
|
+
f.write('PATH = \"/Users/developer/test\"\n')
|
|
927
|
+
p = f.name
|
|
928
|
+
violations = check_forbidden_paths([p])
|
|
929
|
+
os.unlink(p)
|
|
930
|
+
print('SEC-02: Violations caught in .py file:', len(violations))
|
|
931
|
+
assert len(violations) == 0, 'Bypass verified: .py files skipped'
|
|
932
|
+
"
|
|
933
|
+
```
|
|
934
|
+
|
|
935
|
+
#### C. Reproduce QG-01 (Baseline Carte Blanche Suppression)
|
|
936
|
+
```bash
|
|
937
|
+
python3 -c "
|
|
938
|
+
from tools.infer_linter import load_baseline
|
|
939
|
+
b = load_baseline('tools/.infer_baseline.json')
|
|
940
|
+
suppressed = ('tools/core/process_supervisor.py', 'SWALLOWED_EXCEPTION') in b
|
|
941
|
+
print('QG-01: File-wide carte blanche tuple present:', suppressed)
|
|
942
|
+
assert suppressed, 'Flaw verified: Baseline grants whole-file exception immunity'
|
|
943
|
+
"
|
|
944
|
+
```
|
|
945
|
+
|
|
946
|
+
#### D. Reproduce QG-02 (Dummy Bypasses Count)
|
|
947
|
+
```bash
|
|
948
|
+
python3 -c "
|
|
949
|
+
import subprocess
|
|
950
|
+
res = subprocess.run(['git', 'grep', '-n', '-E', r'_\s*=\s*(err|exc|e|fallback_err|False)\b', 'tools/'], capture_output=True, text=True)
|
|
951
|
+
count = len(res.stdout.strip().split('\n'))
|
|
952
|
+
print(f'QG-02: Total dummy bypass assignments in tools/: {count}')
|
|
953
|
+
assert count > 50
|
|
954
|
+
"
|
|
955
|
+
```
|
|
956
|
+
|
|
957
|
+
#### E. Reproduce F-P2-01 (Sub-3000ms Static Sleep Pass-Through)
|
|
958
|
+
```bash
|
|
959
|
+
python3 -c "
|
|
960
|
+
import tempfile, os, subprocess, sys
|
|
961
|
+
with tempfile.TemporaryDirectory() as td:
|
|
962
|
+
fpath = os.path.join(td, 'test_short.py')
|
|
963
|
+
with open(fpath, 'w') as f:
|
|
964
|
+
f.write('import time\ndef test_fn():\n time.sleep(1)\n assert True\n')
|
|
965
|
+
res = subprocess.run([sys.executable, 'tools/infer_linter.py', '--strict', '--dir', td], capture_output=True, text=True)
|
|
966
|
+
print('F-P2-01: time.sleep(1) exit code:', res.returncode)
|
|
967
|
+
assert res.returncode == 0, 'Bypass verified: 1s sleep passed Gate 7'
|
|
968
|
+
"
|
|
969
|
+
```
|
|
970
|
+
|
|
971
|
+
#### F. Reproduce F-P2-02 (Gate 3 Clean POM Semantic Locator Blindspot)
|
|
972
|
+
```bash
|
|
973
|
+
python3 -c "
|
|
974
|
+
import tempfile, os, sys
|
|
975
|
+
sys.path.insert(0, 'tools')
|
|
976
|
+
from verify_all import gate_3_clean_pom
|
|
977
|
+
with tempfile.TemporaryDirectory() as td:
|
|
978
|
+
t_dir = os.path.join(td, 'tests')
|
|
979
|
+
os.makedirs(t_dir)
|
|
980
|
+
with open(os.path.join(t_dir, 'auth.spec.ts'), 'w') as f:
|
|
981
|
+
f.write('test(\"bad pom\", async ({page}) => { await page.getByRole(\"button\").click(); });\n')
|
|
982
|
+
passed = gate_3_clean_pom(td, mode='TARGET')
|
|
983
|
+
print('F-P2-02: Gate 3 passed raw getByRole in test script:', passed)
|
|
984
|
+
assert passed is True, 'Blindspot verified: Playwright semantic locators ignored by Gate 3'
|
|
985
|
+
"
|
|
986
|
+
```
|
|
987
|
+
|
|
988
|
+
#### G. Reproduce QG-03 (Corrupted JSON Output on Stdout)
|
|
989
|
+
```bash
|
|
990
|
+
python3 tools/verify_all.py --format json 2>&1 | python3 -c "
|
|
991
|
+
import sys, json
|
|
992
|
+
raw = sys.stdin.read()
|
|
993
|
+
try:
|
|
994
|
+
json.loads(raw)
|
|
995
|
+
print('Valid JSON')
|
|
996
|
+
except json.JSONDecodeError as e:
|
|
997
|
+
print('QG-03: JSON parsing failed as predicted:', e)
|
|
998
|
+
"
|
|
999
|
+
```
|
|
1000
|
+
|
|
1001
|
+
#### H. Reproduce F-P7-01 (`project_state.py status` Crash)
|
|
1002
|
+
```bash
|
|
1003
|
+
python3 -c "
|
|
1004
|
+
import subprocess, sys
|
|
1005
|
+
res = subprocess.run([sys.executable, 'tools/project_state.py', 'status'], capture_output=True, text=True)
|
|
1006
|
+
print('F-P7-01: Return code:', res.returncode)
|
|
1007
|
+
print('Error output snippet:', res.stderr.strip().split('\n')[-1])
|
|
1008
|
+
assert res.returncode != 0 and 'FileNotFoundError' in res.stderr
|
|
1009
|
+
"
|
|
1010
|
+
```
|
|
1011
|
+
|
|
1012
|
+
#### I. Reproduce ROUTER-01 (Depth-2 Monorepo Marker Truncation)
|
|
1013
|
+
```bash
|
|
1014
|
+
python3 -c "
|
|
1015
|
+
import tempfile, os
|
|
1016
|
+
from tools.rules_router import detect_stacks
|
|
1017
|
+
with tempfile.TemporaryDirectory() as td:
|
|
1018
|
+
deep = os.path.join(td, 'pkg', 'subpkg', 'deep')
|
|
1019
|
+
os.makedirs(deep)
|
|
1020
|
+
with open(os.path.join(deep, 'playwright.config.ts'), 'w') as f:
|
|
1021
|
+
f.write('export default {};\n')
|
|
1022
|
+
stacks = detect_stacks(td)
|
|
1023
|
+
print('ROUTER-01: Detected stacks:', stacks)
|
|
1024
|
+
assert stacks == ['unknown'], 'Truncation verified: Marker at depth 3 missed'
|
|
1025
|
+
"
|
|
1026
|
+
```
|
|
1027
|
+
|
|
1028
|
+
---
|
|
1029
|
+
|
|
1030
|
+
### 5.3 Audit Invalidation Conditions
|
|
1031
|
+
This audit's conclusions and ratings would be invalidated if:
|
|
1032
|
+
1. `tests/test_benchmark_context.py` fails when run against `benchmarks/CONTEXT_TOKEN_BENCHMARK.md`.
|
|
1033
|
+
2. `tools/rules_router.py` returns non-zero token savings when executed against an empty directory or `/tmp`.
|
|
1034
|
+
3. `tools/distill_log.py` emits `ALL TESTS PASSED` when passed a non-zero exit code or fails to raise `ValueError` on negative token budgets.
|
|
1035
|
+
4. `tools/micro_mutator.py` fails to restore pristine code after dynamic test execution or allows tautological tests (`assert True`) to achieve a passing score.
|
|
1036
|
+
5. `tools/reversible_sandbox.py` clears `.synapse/undo_wal.jsonl` when an undo command returns exit code 1.
|
|
1037
|
+
6. UTF-8 BOM bytes (`\xef\xbb\xbf`) are detected in any candidate repository file.
|
|
1038
|
+
7. `tools/author_integrity.py` passes when canonical author name `Thế Vị` or handle `@TeeVee98` is stripped from `LICENSE` or `package.json`.
|