@devran-ai/kit 4.1.0

package/.agent/commands/status.md

@@ -0,0 +1,59 @@
+---
+name: status
+description: Display current session status and context
+workflow: status
+---
+
+# /status
+
+> **Purpose**: Quick overview of current session state
+
+---
+
+## Usage
+
+```
+/status
+```
+
+---
+
+## Behavior
+
+1. **Load Session State** from `session-state.json`
+2. **Check Git Status** - Current branch, uncommitted changes
+3. **Check Build Status** - Last build result
+4. **Display Context** - Current task, open items
+
+---
+
+## Output
+
+```markdown
+## Session Status
+
+**Branch**: feature/auth
+**Last Commit**: abc1234 - feat: add login form
+**Uncommitted Changes**: 3 files
+
+## Current Task
+
+[Current task description or "No active task"]
+
+## Open Items
+
+- [x] Item 1
+- [ ] Item 2
+
+## Quick Actions
+
+- `/plan` - Create new plan
+- `/verify` - Run verification
+```
+
+---
+
+## Related Commands
+
+- `/checkpoint` - Save current state
+- `/compact` - Manage context

package/.agent/commands/tdd.md

@@ -0,0 +1,73 @@
+---
+name: tdd
+description: Start test-driven development workflow for a component
+invokes: tdd-guide
+workflow: test
+---
+
+# /tdd
+
+> **Purpose**: Enforce test-first development with RED-GREEN-REFACTOR cycle
+
+---
+
+## Usage
+
+```
+/tdd [component/feature name]
+```
+
+## Examples
+
+```
+/tdd UserService
+/tdd AuthController login endpoint
+/tdd calculateTotal utility function
+```
+
+---
+
+## Behavior
+
+1. **Invoke TDD Guide Agent**
+2. **Analyze Component** - Understand what to test
+3. **Write Failing Test** (RED) - Create test that fails
+4. **Verify Failure** - Run test, confirm it fails
+5. **Write Implementation** (GREEN) - Minimal code to pass
+6. **Verify Pass** - Run test, confirm it passes
+7. **Refactor** (IMPROVE) - Clean up while keeping tests green
+8. **Check Coverage** - Verify 80%+ achieved
+
+---
+
+## Output
+
+```markdown
+## TDD Session: UserService
+
+### RED Phase 🔴
+
+Created test: `user.service.spec.ts`
+Test status: FAILING ✓
+
+### GREEN Phase 🟢
+
+Implementation: `user.service.ts`
+Test status: PASSING ✓
+
+### Coverage
+
+- Statements: 92%
+- Branches: 85%
+- Functions: 90%
+- Lines: 92%
+
+Status: ✅ TDD Complete
+```
+
+---
+
+## Related Commands
+
+- `/verify` - Full verification loop
+- `/code-review` - Review the implementation

package/.agent/commands/verify.md

@@ -0,0 +1,58 @@
+---
+name: verify
+description: Run full verification loop (build, lint, test, coverage)
+invokes: verification-loop skill
+---
+
+# /verify
+
+> **Purpose**: Comprehensive quality verification before commit/deploy
+
+---
+
+## Usage
+
+```
+/verify
+/verify --fix  # Auto-fix lint issues
+```
+
+---
+
+## Behavior
+
+1. **Build Check** - Compile TypeScript
+2. **Lint Check** - ESLint/Prettier
+3. **Type Check** - Strict TypeScript
+4. **Unit Tests** - All unit tests
+5. **Integration Tests** - API tests
+6. **Coverage Check** - 80%+ required
+7. **Security Scan** - Basic vulnerability check
+
+---
+
+## Output
+
+```markdown
+## Verification Report
+
+| Check      | Status | Time |
+| :--------- | :----- | :--- |
+| Build      | ✅     | 2.3s |
+| Lint       | ✅     | 1.2s |
+| Types      | ✅     | 3.1s |
+| Unit Tests | ✅     | 8.5s |
+| Coverage   | ✅ 92% | -    |
+| Security   | ✅     | 1.0s |
+
+**Overall**: ✅ PASS
+
+Ready to commit!
+```
+
+---
+
+## Related Commands
+
+- `/security-scan` - Deep security audit
+- `/code-review` - Manual code review

package/.agent/contexts/brainstorm.md

@@ -0,0 +1,26 @@
+---
+mode: brainstorm
+description: "Ideas-only exploration mode — Socratic discovery without implementation"
+constraints: [no-code, no-modifications, socratic-only]
+agent-focus: [planner, explorer-agent]
+---
+
+# Brainstorm Mode Context
+
+## Behavioral Rules
+
+- Generate ideas and alternatives only — never write code
+- Ask Socratic questions to clarify requirements
+- Present minimum 3 options for every decision point
+- Support each option with evidence or industry precedent
+- Never commit to a single approach without user confirmation
+
+## Loaded Skills
+
+- `brainstorming` — Socratic discovery protocol
+
+## Exit Criteria
+
+- Requirements are clarified with concrete acceptance criteria
+- At least 3 viable approaches documented with trade-offs
+- User selects preferred approach for planning phase

package/.agent/contexts/debug.md

@@ -0,0 +1,28 @@
+---
+mode: debug
+description: "Systematic investigation mode — minimal changes, preserve evidence"
+constraints: [evidence-first, minimal-changes, root-cause-required]
+agent-focus: [build-error-resolver, tdd-guide]
+---
+
+# Debug Mode Context
+
+## Behavioral Rules
+
+- Identify root cause before applying any fix
+- Preserve all evidence (logs, stack traces, reproduction steps)
+- Make minimal, surgical changes — one fix per iteration
+- Always verify the fix resolves the original issue
+- Add regression tests to prevent recurrence
+
+## Loaded Skills
+
+- `debugging-strategies` — Systematic hypothesis testing
+- `testing-patterns` — Regression test creation
+
+## Exit Criteria
+
+- Root cause identified and documented
+- Fix applied with minimal code changes
+- Regression test added and passing
+- No new issues introduced (full test suite passes)

package/.agent/contexts/implement.md

@@ -0,0 +1,29 @@
+---
+mode: implement
+description: "Code-focused execution mode — follow approved plan, build incrementally"
+constraints: [plan-required, incremental-commits, tests-mandatory]
+agent-focus: [architect, tdd-guide, code-reviewer]
+---
+
+# Implement Mode Context
+
+## Behavioral Rules
+
+- Follow the approved implementation plan step-by-step
+- Write code that passes existing tests and adds new ones
+- Commit incrementally — one logical change per commit
+- Do not deviate from the plan without explicit user approval
+- Apply project conventions and coding standards strictly
+
+## Loaded Skills
+
+- `clean-code` — Code quality principles
+- `testing-patterns` — TDD and test structure
+- `architecture` — System design patterns
+
+## Exit Criteria
+
+- All plan steps completed
+- All tests pass (existing + new)
+- Build succeeds without errors
+- Code reviewed for quality standards

package/.agent/contexts/plan-quality-log.md

@@ -0,0 +1,30 @@
+# Plan Quality Log
+
+> Persistent record of plan accuracy and learnings.
+> Read by the planner agent at planning time to improve future plans.
+> Updated by the plan-retrospective after each planned task completes.
+
+---
+
+## Log
+
+| Date | Plan | Quality Score | Files Predicted | Files Actual | Surprises | Estimate Drift | Key Learning |
+|------|------|--------------|-----------------|--------------|-----------|----------------|--------------|
+
+---
+
+## Aggregate Insights
+
+> This section is updated periodically to summarize patterns across multiple retrospectives.
+
+### Common Surprise File Patterns
+
+*No data yet — will populate after first retrospective.*
+
+### Estimate Drift Trends
+
+*No data yet — will populate after first retrospective.*
+
+### Risk Prediction Accuracy
+
+*No data yet — will populate after first retrospective.*

package/.agent/contexts/review.md

@@ -0,0 +1,27 @@
+---
+mode: review
+description: "Quality-focused critical analysis mode — verify standards compliance"
+constraints: [critical-lens, no-bias, evidence-based]
+agent-focus: [code-reviewer, security-reviewer]
+---
+
+# Review Mode Context
+
+## Behavioral Rules
+
+- Apply critical lens — assume nothing is correct until verified
+- Check against all quality gates (lint, types, tests, security)
+- No defense bias — report issues objectively regardless of author
+- Classify findings by severity (CRITICAL / HIGH / MEDIUM / LOW)
+- Block deployment for any CRITICAL or HIGH severity issue
+
+## Loaded Skills
+
+- `verification-loop` — Sequential quality gate pipeline
+- `security-practices` — OWASP vulnerability checks
+
+## Exit Criteria
+
+- All quality gates pass (lint, type-check, tests, security)
+- All CRITICAL and HIGH issues resolved
+- Review report generated with verdict (APPROVE / BLOCK / WARNING)

package/.agent/contexts/ship.md

@@ -0,0 +1,28 @@
+---
+mode: ship
+description: "Deployment-focused risk-aware mode — pre-flight checks, monitoring, rollback readiness"
+constraints: [rollback-plan-required, health-check-mandatory, no-friday-deploys]
+agent-focus: [devops-engineer, reliability-engineer]
+---
+
+# Ship Mode Context
+
+## Behavioral Rules
+
+- Verify all pre-flight checks pass before any deployment
+- Require a documented rollback plan for every deploy
+- Monitor health checks for minimum 15 minutes post-deploy
+- No deployments on Fridays unless critical hotfix
+- Follow the Production Merge Discipline from global rules
+
+## Loaded Skills
+
+- `deployment-procedures` — CI/CD workflows and rollback strategies
+
+## Exit Criteria
+
+- All pre-flight checks pass (tests, build, security, env vars)
+- Rollback plan documented and verified
+- Deployment executed successfully
+- Health checks pass for observation period
+- No error spikes in monitoring

package/.agent/decisions/001-trust-grade-governance.md

@@ -0,0 +1,46 @@
+# ADR-001: Trust-Grade Governance Model
+
+## Status
+
+**Accepted** — 2025-12-01
+
+## Context
+
+Devran AI Kit needed a governance philosophy that would ensure AI-assisted development remains safe, predictable, and aligned with engineering best practices. Without formal constraints, AI agents may produce technically correct but architecturally misaligned solutions, bypass security practices, or make decisions that optimize for speed over stability.
+
+## Decision
+
+We adopted the **Trust-Grade Governance Model** with the following immutable constraints:
+
+| Priority | Principle | Meaning |
+|:---------|:----------|:--------|
+| 1 | Trust > Speed | Never sacrifice user trust for faster delivery |
+| 2 | Completion > Suggestion | Finish current work before proposing new work |
+| 3 | Consistency > Speed | Update all affected files, not just the target |
+| 4 | Explainability > Performance | Understandable decisions beat clever shortcuts |
+| 5 | Safety > Growth | Never compromise security, data privacy, or stability |
+
+These principles are enforced through:
+- **rules.md**: The binding identity contract loaded at every session
+- **Session management**: State tracked across sessions via `session-context.md` and `session-state.json`
+- **Checklists**: Pre/post session quality gates
+- **Hooks**: Event-driven automation for enforcement
+
+## Consequences
+
+### Positive
+- Consistent AI behavior across sessions and projects
+- Security-by-default through principle ordering
+- Predictable decision-making under ambiguity
+- Maintainable codebase through Consistency > Speed
+
+### Negative
+- Slower initial response time (context loading)
+- Higher token usage per session (governance overhead)
+- May feel restrictive for rapid prototyping scenarios
+
+## Alternatives Considered
+
+1. **No governance** — Pure LLM default behavior. Rejected due to inconsistency and safety risks.
+2. **Permission-based** — Ask user for every decision. Rejected due to excessive friction.
+3. **Role-based** — Different governance per agent. Rejected due to complexity of maintaining cross-agent consistency.

package/.agent/decisions/002-cross-ide-generation.md

@@ -0,0 +1,15 @@
+# ADR-002: Cross-IDE Configuration Generation
+
+## Status: Accepted
+
+## Context
+Kit v4.0.0 only supported Claude Code via `.agent/`. Users of Cursor, OpenCode, and Codex had no native integration.
+
+## Decision
+Generate IDE-specific configs at `kit init` time from `manifest.json` (single source of truth). Configs live outside `.agent/` in native IDE directories.
+
+## Consequences
+- One command configures all IDEs
+- Configs stay in sync with manifest
+- Non-destructive: existing IDE dirs are preserved unless --force
+- IDE configs are gitignored by default (generated, not source)

package/.agent/engine/identity.json

@@ -0,0 +1,4 @@
+{
+  "developers": [],
+  "activeId": null
+}

package/.agent/engine/loading-rules.json

@@ -0,0 +1,193 @@
+{
+  "schemaVersion": "1.0.0",
+  "defaultLoad": [
+    "rules.md",
+    "manifest.json",
+    "engine/workflow-state.json"
+  ],
+  "sessionLoad": [
+    "session-context.md",
+    "session-state.json"
+  ],
+  "domainRules": [
+    {
+      "domain": "security",
+      "keywords": ["auth", "security", "vulnerability", "password", "jwt", "oauth", "cors", "xss", "csrf", "injection"],
+      "implicitTriggers": [
+        "login", "signup", "register", "form", "payment", "checkout",
+        "user data", "profile", "upload", "download", "admin", "dashboard",
+        "permission", "role", "token", "session", "cookie"
+      ],
+      "loadAgents": ["security-reviewer"],
+      "loadSkills": ["security-practices"]
+    },
+    {
+      "domain": "testing",
+      "keywords": ["test", "coverage", "e2e", "playwright", "jest", "vitest", "tdd", "unit test", "integration test"],
+      "loadAgents": ["tdd-guide", "e2e-runner"],
+      "loadSkills": ["testing-patterns", "webapp-testing"]
+    },
+    {
+      "domain": "architecture",
+      "keywords": ["design", "structure", "pattern", "ddd", "hexagonal", "layer", "module", "monolith", "microservice"],
+      "loadAgents": ["architect"],
+      "loadSkills": ["architecture", "clean-code"]
+    },
+    {
+      "domain": "frontend",
+      "keywords": ["react", "next.js", "vue", "component", "css", "styling", "responsive", "a11y", "accessibility", "ui", "ux"],
+      "implicitTriggers": ["button", "modal", "form field", "dropdown", "navigation bar", "sidebar", "theme", "dark mode", "animation", "transition", "layout", "grid", "flexbox", "tailwind", "styled"],
+      "loadAgents": ["frontend-specialist"],
+      "loadSkills": ["frontend-patterns", "mobile-design"]
+    },
+    {
+      "domain": "backend",
+      "keywords": ["api", "server", "node", "nestjs", "express", "middleware", "rest", "graphql", "endpoint"],
+      "implicitTriggers": ["route", "controller", "service layer", "webhook", "queue", "worker", "cron", "batch", "rate limit", "pagination", "validation"],
+      "loadAgents": ["backend-specialist"],
+      "loadSkills": ["api-patterns", "nodejs-patterns"]
+    },
+    {
+      "domain": "database",
+      "keywords": ["database", "sql", "postgresql", "prisma", "migration", "schema", "query", "orm", "redis"],
+      "implicitTriggers": ["table", "column", "index", "foreign key", "transaction", "connection pool", "backup", "replica", "sharding", "partition"],
+      "loadAgents": ["database-architect"],
+      "loadSkills": ["database-design"]
+    },
+    {
+      "domain": "devops",
+      "keywords": ["deploy", "ci", "cd", "docker", "kubernetes", "railway", "vercel", "github actions", "pipeline", "terraform", "gitops", "canary", "helm"],
+      "implicitTriggers": ["environment variable", "container", "image", "registry", "rollback", "blue-green", "feature flag", "infrastructure"],
+      "loadAgents": ["devops-engineer"],
+      "loadSkills": ["docker-patterns", "deployment-procedures"]
+    },
+    {
+      "domain": "performance",
+      "keywords": ["slow", "optimize", "speed", "bundle", "lighthouse", "web vitals", "lcp", "cls", "fid", "cache", "cdn", "latency", "p99", "tracing"],
+      "loadAgents": ["performance-optimizer"],
+      "loadSkills": ["performance-profiling"]
+    },
+    {
+      "domain": "reliability",
+      "keywords": ["reliability", "uptime", "monitoring", "sre", "sla", "slo", "sli", "incident", "chaos", "resilience", "error-budget", "golden-signals", "on-call", "capacity"],
+      "loadAgents": ["reliability-engineer"],
+      "loadSkills": []
+    },
+    {
+      "domain": "observability",
+      "keywords": ["logging", "tracing", "metrics", "monitoring", "alerting", "opentelemetry", "grafana", "prometheus", "datadog", "observability", "dashboard"],
+      "loadAgents": ["reliability-engineer", "devops-engineer"],
+      "loadSkills": []
+    },
+    {
+      "domain": "mobile",
+      "keywords": ["mobile", "react native", "expo", "ios", "android", "app store", "native"],
+      "loadAgents": ["mobile-developer"],
+      "loadSkills": ["mobile-design"]
+    },
+    {
+      "domain": "documentation",
+      "keywords": ["docs", "readme", "document", "api docs", "jsdoc", "changelog", "adr", "decision record", "knowledge"],
+      "loadAgents": ["doc-updater", "knowledge-agent"],
+      "loadSkills": []
+    },
+    {
+      "domain": "planning",
+      "keywords": ["plan", "roadmap", "sprint", "milestone", "task", "breakdown", "estimate"],
+      "loadAgents": ["planner", "sprint-orchestrator"],
+      "loadSkills": ["plan-writing", "brainstorming"]
+    },
+    {
+      "domain": "debugging",
+      "keywords": ["bug", "error", "fix", "crash", "exception", "stack trace", "not working"],
+      "loadAgents": ["build-error-resolver"],
+      "loadSkills": ["debugging-strategies"]
+    },
+    {
+      "domain": "refactoring",
+      "keywords": ["refactor", "clean", "improve", "dead code", "technical debt", "simplify"],
+      "loadAgents": ["refactor-cleaner"],
+      "loadSkills": ["clean-code"]
+    },
+    {
+      "domain": "typescript",
+      "keywords": ["typescript", "ts", "type safety", "tsconfig", "tsc", "generics", "type guard", "discriminated union"],
+      "loadAgents": ["typescript-reviewer"],
+      "loadSkills": ["typescript-expert"]
+    },
+    {
+      "domain": "python",
+      "keywords": ["python", "django", "flask", "fastapi", "pep8", "mypy", "pytest", "asyncio", "pydantic"],
+      "loadAgents": ["python-reviewer"],
+      "loadSkills": []
+    },
+    {
+      "domain": "golang",
+      "keywords": ["go", "golang", "goroutine", "channel", "context", "errgroup", "go mod", "staticcheck"],
+      "loadAgents": ["go-reviewer"],
+      "loadSkills": []
+    }
+  ],
+  "contextBudget": {
+    "maxAgentsPerSession": 4,
+    "maxSkillsPerSession": 8,
+    "warningThresholdPercent": 80,
+    "description": "Limits prevent context overflow. The kit selects only domain-relevant agents and skills for each task, keeping within the LLM context window."
+  },
+  "workflowBindings": [
+    { "workflow": "brainstorm", "loadAgents": [], "loadSkills": ["brainstorming"], "bindingType": "inferred" },
+    { "workflow": "quality-gate", "loadAgents": [], "loadSkills": ["brainstorming"], "loadRules": ["quality-gate"], "bindingType": "inferred" },
+    { "workflow": "plan", "loadAgents": ["planner"], "loadSkills": ["plan-writing", "brainstorming", "plan-validation"], "bindingType": "explicit" },
+    { "workflow": "create", "loadAgents": [], "loadSkills": ["app-builder", "clean-code"], "bindingType": "inferred" },
+    { "workflow": "enhance", "loadAgents": [], "loadSkills": ["clean-code", "testing-patterns"], "bindingType": "inferred" },
+    { "workflow": "preview", "loadAgents": [], "loadSkills": ["shell-conventions"], "bindingType": "inferred" },
+    { "workflow": "ui-ux-pro-max", "loadAgents": [], "loadSkills": ["ui-ux-pro-max", "frontend-patterns", "mobile-design"], "bindingType": "inferred" },
+    { "workflow": "test", "loadAgents": [], "loadSkills": ["testing-patterns", "webapp-testing"], "bindingType": "inferred" },
+    { "workflow": "review", "loadAgents": [], "loadSkills": ["verification-loop"], "bindingType": "inferred" },
+    { "workflow": "pr", "loadAgents": ["pr-reviewer"], "loadSkills": ["git-workflow", "pr-toolkit", "verification-loop"], "bindingType": "explicit" },
+    { "workflow": "pr-review", "loadAgents": ["pr-reviewer"], "loadSkills": ["pr-toolkit", "verification-loop"], "bindingType": "explicit" },
+    { "workflow": "pr-fix", "loadAgents": ["pr-reviewer"], "loadSkills": ["pr-toolkit", "verification-loop"], "bindingType": "explicit" },
+    { "workflow": "pr-merge", "loadAgents": ["pr-reviewer"], "loadSkills": ["pr-toolkit", "verification-loop"], "bindingType": "explicit" },
+    { "workflow": "pr-split", "loadAgents": ["pr-reviewer"], "loadSkills": ["pr-toolkit"], "bindingType": "explicit" },
+    { "workflow": "deploy", "loadAgents": [], "loadSkills": ["deployment-procedures"], "bindingType": "inferred" },
+    { "workflow": "debug", "loadAgents": [], "loadSkills": ["debugging-strategies"], "bindingType": "inferred" },
+    {
+      "workflow": "orchestrate",
+      "loadAgents": ["planner", "explorer-agent"],
+      "loadSkills": ["parallel-agents", "intelligent-routing"],
+      "bindingType": "explicit",
+      "dynamicAgentPool": ["architect", "backend-specialist", "frontend-specialist", "mobile-developer", "database-architect", "security-reviewer", "tdd-guide", "e2e-runner", "devops-engineer", "performance-optimizer", "reliability-engineer", "refactor-cleaner", "code-reviewer", "typescript-reviewer", "python-reviewer", "go-reviewer"]
+    },
+    { "workflow": "retrospective", "loadAgents": [], "loadSkills": ["verification-loop"], "bindingType": "inferred" },
+    { "workflow": "status", "loadAgents": [], "loadSkills": ["verification-loop"], "bindingType": "inferred" }
+  ],
+  "planningMandates": {
+    "description": "Mandatory resources loaded for every /plan invocation regardless of keyword matching. Ensures cross-cutting concerns are never omitted.",
+    "alwaysLoadRules": ["security", "testing", "coding-style", "documentation", "architecture"],
+    "alwaysLoadSkills": ["security-practices", "testing-patterns"],
+    "crossCuttingSections": [
+      "security-considerations",
+      "testing-strategy",
+      "rollback-strategy",
+      "documentation-updates"
+    ],
+    "specialistContributors": {
+      "architect": {
+        "contributes": "architecture-impact",
+        "triggerWhen": "medium-or-large-task"
+      },
+      "security-reviewer": {
+        "contributes": "threat-assessment",
+        "triggerWhen": "medium-or-large-task",
+        "crossCuttingAlways": true,
+        "note": "Cross-cutting security section (Tier 1) is always required via rule consultation. Full specialist invocation with STRIDE model is for Medium+ tasks."
+      },
+      "tdd-guide": {
+        "contributes": "test-strategy",
+        "triggerWhen": "medium-or-large-task",
+        "crossCuttingAlways": true,
+        "note": "Cross-cutting testing section (Tier 1) is always required via rule consultation. Full specialist invocation with test matrix is for Medium+ tasks."
+      }
+    }
+  }
+}

package/.agent/engine/marketplace-index.json

@@ -0,0 +1,29 @@
+{
+  "entries": [
+    {
+      "name": "ag-skill-docker-compose",
+      "description": "Docker Compose orchestration skill for multi-container deployments",
+      "repository": "https://github.com/devran-ai/ag-skill-docker-compose",
+      "version": "1.0.0",
+      "tags": ["docker", "compose", "containers", "deployment", "devops"],
+      "author": "devran-ai"
+    },
+    {
+      "name": "ag-skill-graphql-designer",
+      "description": "GraphQL schema design and query generation skill",
+      "repository": "https://github.com/devran-ai/ag-skill-graphql-designer",
+      "version": "0.9.0",
+      "tags": ["graphql", "api", "schema", "design", "queries"],
+      "author": "devran-ai"
+    },
+    {
+      "name": "ag-skill-accessibility-audit",
+      "description": "WCAG compliance auditing and remediation suggestions",
+      "repository": "https://github.com/devran-ai/ag-skill-accessibility-audit",
+      "version": "1.1.0",
+      "tags": ["accessibility", "wcag", "a11y", "audit", "compliance"],
+      "author": "devran-ai"
+    }
+  ],
+  "lastUpdated": null
+}

package/.agent/engine/mcp-servers/filesystem.json

@@ -0,0 +1,9 @@
+{
+  "name": "filesystem",
+  "description": "Local filesystem — read, write, search, directory listing",
+  "transport": "stdio",
+  "command": "npx",
+  "args": ["-y", "@modelcontextprotocol/server-filesystem", "${PROJECT_ROOT}"],
+  "env": {},
+  "capabilities": ["read", "write", "search", "directory-listing"]
+}