This release includes 1 security fix for security teams reviewing exposed deployments.
Topics
+10 more
Affected surfaces
Summary
AI summaryAdded MCP structured content and shared schemas for type‑safe AI model consumption.
Full changelog
Added
- MCP Structured Content — all 17 MCP tools now return
structuredContentwith ZodoutputSchemafor type-safe AI model consumption - Shared MCP schemas — reusable Zod schemas in
src/mcp/schemas/(audit, common, health, server) - MCP SDK round-trip tests — 26 tests verifying
normalizeObjectSchema+safeParseAsyncfor all 17 outputSchemas - Plugin manifest —
.claude-plugin/plugin.jsonwith mcpServers, skills, hooks for Claude Code plugin distribution
Fixed
- outputSchema wrap pattern —
z.discriminatedUnionwrapped inz.object({ result })to satisfy MCP SDK'snormalizeObjectSchema - Plugin manifest paths —
./prefix required for CC plugin validator;agentsfield removed (unsupported) - ip-address XSS vulnerability — override to ^10.1.1 (dependabot #9)
Changed
- P131 code quality sweep — catch blocks with debugLog, console.log→logger migration
- P130 file split —
interactive.ts→ 8 modules,lock.ts→ 8 modules
Security Fixes
- ip-address XSS vulnerability fixed; dependency upgraded to ^10.1.1
Weekly OSS security release digest.
The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.
No spam, unsubscribe anytime.
Share this release
About kastelldev/kastell
Server security auditing and hardening toolkit. 413 security checks across 29 categories (SSH, Firewall, Docker, TLS, HTTP Headers), CIS/PCI-DSS/HIPAA compliance mapping, 19-step production hardening, fleet management, and forensic evidence collection. Supports Hetzner, DigitalOcean, Vultr, and Linode. 13 MCP tools.
Related context
Related tools
Beta — feedback welcome: [email protected]