obsidian-plugin/aaronsb_obsidian-mcp-plugin
1
0
Fork 0
mirror of https://github.com/aaronsb/obsidian-mcp-plugin.git synced 2026-07-22 06:45:14 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions
aaronsb_obsidian-mcp-plugin/CHANGELOG.md
Aaron Bockelie f3dac1847e feat(vault.read): faithful-by-default reads with char-budget pagination (ADR-203, #133) 
Default vault.read no longer fragments-and-flattens. New contract:
- fits READ_PAGE_CHARS (50000) → whole file, byte-exact, one load (common case)
- exceeds → verbatim page 1, single contiguous block, with line bookends
  (lineStart/lineEnd/totalLines + nextPage); page=N to continue. Absolute
  line numbers preserved so edit.at_line still works on large files.
- returnFullFile:true → entire file verbatim (explicit large override; param
  retained & repurposed, not retired)
- query/strategy/maxFragments → semantic fragments (unchanged)
- structured envelope no longer double-encodes the body (metadata sans body)

Budget is char-based on purpose: line count is an invalid proxy for context
cost; only bookends are line-based (for at_line). Hard invariant: a default
read must never hand the agent a context-breaking raw dump.

Also fixes the latent formatFileRead crash (_Formatter error_) on full-file
shapes; formatter now renders verbatim + a Pagination section.

src/utils/file-reader.ts rewritten; vault.ts threads `page`; formatter
hardened; tool description + CHANGELOG (breaking) updated.
make check green (0 errors, baseline 5 warnings, 243/243 incl. 8 new
ADR-203 round-trip/pagination/fidelity tests).
2026-05-19 00:30:31 -05:00

9.7 KiB
Raw Blame History

Changelog

All notable changes to the Obsidian MCP Plugin will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[Unreleased]

Changed

  • ⚠️ BREAKING: vault.read is now faithful by default (ADR-203, #133). It returns the complete, byte-exact file source (no more newline-flattened fragments) when the file fits a ~50k-char budget. Large files return a verbatim page 1 with absolute line bookends (page=N to continue) instead of a context-breaking raw dump. returnFullFile: true is the explicit whole-large-file override; query/strategy/maxFragments still return semantic fragments. The structured envelope no longer double-encodes the body. Clients that relied on the old fragmented default should pass fragment params explicitly.

Security

  • 🔴 CRITICAL: Identified authentication vulnerability - no API key validation (#9)
  • 🔴 CRITICAL: Identified path traversal vulnerability in file operations (#10)
  • 🟠 HIGH: Identified missing input validation across all operations (#11)
  • 🟠 HIGH: Identified insecure session management implementation (#12)

[0.9.3] - 2025-01-19

Added

  • Enhanced Tag Search: Search now includes frontmatter tags in addition to content tags (thanks @lukemt!) (#28)
  • Dynamic HTTPS Configuration: Configuration examples automatically adapt based on certificate type
    • NODE_TLS_REJECT_UNAUTHORIZED only shown when using self-signed certificates
    • Informational notes guide users based on their certificate configuration
  • Clearer Server Settings: Separate "Enable HTTP Server" and "Enable HTTPS Server" toggles
    • At least one protocol must always be enabled
    • Dynamic descriptions explain when each can be disabled
    • Status bar shows active protocols (HTTP/HTTPS/both)

Fixed

  • Windows Configuration: Added documentation for Windows-specific command line parsing issues (#30)
  • Port Conflict Detection: Fixed false detection of own running server as a conflict
    • Port availability only checked when changing ports or starting server
    • No longer shows confusing "port in use" messages for running server

Changed

  • Security policy (SECURITY.md) for vulnerability reporting
  • Contributing guidelines (CONTRIBUTING.md)
  • Issue templates for bug reports and feature requests
  • GitHub labels for security, priority, and technical debt tracking
  • Comprehensive security audit documentation
  • Project structure improved with proper open-source documentation

[0.5.14] - 2025-01-11

Added

  • Advanced File Operations: New vault actions for complex file manipulation
    • split - Split files into multiple parts with 4 strategies:
      • heading - Split by markdown heading levels
      • delimiter - Split by custom delimiter string
      • lines - Split by line count per file
      • size - Split by character count with smart word boundaries
    • combine - Merge multiple files into one with options:
      • Custom separators between files
      • Optional filename headers
      • Sort files before combining (by name/size/date)
    • concatenate - Simple two-file joining (append/prepend/new)
    • All operations include semantic workflow hints
    • Smart content preservation in split operations

[0.5.13] - 2025-01-10

Added

  • File Management Operations: New vault actions for organizing files
    • move - Move files to new locations with automatic link updates
    • rename - Rename files in place with automatic link updates
    • copy - Create copies of files with optional overwrite
    • All operations include semantic workflow hints for next actions
    • Uses native Obsidian file manager when available for link preservation
    • Fallback to copy/delete for environments without direct API access

[0.5.12] - 2025-01-07

Changed

  • Enhanced Tool Descriptions: Improved clarity for AI agents
    • Added single emoji per operation for visual categorization (📁✏️👁️💡ℹ️🕸️)
    • Clarified action descriptions with specific use cases
    • Refined parameter descriptions to avoid ambiguity
    • Better disambiguation between similar actions (e.g., window vs file)
    • Balanced approach: ~17% character increase for significant clarity gains

Design Philosophy

  • Single emoji per operation category to aid visual scanning
  • No emojis in parameters to avoid style contamination in user content
  • Trust the semantic hinting layer for workflow guidance
  • Focus on what makes each action unique

[0.5.11] - 2025-07-07

Added

  • Structured Patch Targeting: Precise document modifications
    • Target headings with targetType: 'heading' and nested paths like "Section::Subsection"
    • Target blocks with targetType: 'block' using block IDs (^blockId)
    • Target frontmatter with targetType: 'frontmatter' for field updates
    • All modes support append, prepend, and replace operations

Technical Implementation

  • New helper methods: patchHeading(), patchBlock(), patchFrontmatter()
  • Intelligent section boundary detection for heading operations
  • Automatic frontmatter creation if none exists
  • Maintains exact whitespace and formatting of untargeted content
  • 7 comprehensive unit tests covering all patch modes

Fixed

  • Patch operations now properly handle structured targeting as originally designed
  • Resolved disconnect between semantic API design and implementation

[0.5.10] - 2025-07-07

Fixed

  • Patch Operation Silent Failures: Resolved parameter mismatch issue
    • Patch operations were returning success without modifying files
    • Fixed parameter passing between semantic router and API layer
    • The router now correctly maps oldText/newText to old_text/new_text
    • Resolves issue #4 where patch operations failed silently

Changed

  • Updated semantic router to properly pass patch parameters
  • Maintained backward compatibility with existing patch operations

[0.5.9] - 2025-07-05

Added

  • True Concurrent Sessions Support: Multiple AI agents can work simultaneously without blocking
    • Session-isolated MCP server pool architecture
    • Each session gets its own complete MCP server instance
    • Automatic session management with 1-hour timeout
    • Session reuse for reconnecting clients
  • Session Monitoring: New obsidian://session-info resource shows active sessions
  • Enhanced Documentation:
    • mcp-remote configuration for Claude Desktop
    • Dynamic resource count in plugin settings
    • Both direct HTTP and mcp-remote options documented

Fixed

  • Concurrent sessions now truly run in parallel without interference
  • Graph traversal operations no longer block other sessions
  • Session context properly isolated between connections
  • Plugin settings UI shows all available resources

Changed

  • Moved concurrency isolation to a higher architectural level
  • MCP SDK remains unaware of concurrency (simpler, cleaner design)
  • Transparent request routing to session-specific servers

[0.5.8c] - 2025-07-05

Added

  • Session-isolated MCP server pool architecture for true concurrent processing
  • MCPServerPool class to manage multiple isolated server instances
  • Server pool statistics in session-info resource

Fixed

  • Concurrent sessions now truly run in parallel without interference
  • Graph traversal operations no longer block other sessions
  • Session context properly isolated between connections

Changed

  • Moved concurrency isolation to a higher architectural level
  • Each session gets its own complete MCP server instance
  • MCP SDK remains unaware of concurrency (simpler, cleaner design)
  • Transparent request routing to session-specific servers

[0.5.8b] - 2025-07-05

Added

  • Worker thread infrastructure for CPU-intensive operations
  • WorkerManager for session-based worker lifecycle
  • Semantic worker for processing search and graph operations

Changed

  • Updated build process to compile worker scripts
  • Operations can be offloaded to worker threads

Fixed

  • Attempted to resolve concurrent session blocking (partial fix)

[0.5.8a] - 2025-07-05

Added

  • Concurrent Sessions Support: Multiple AI agents can now work simultaneously

    • Session-based connection pooling with up to 32 concurrent operations
    • Each MCP client gets a unique session ID for isolation
    • Session tracking and automatic cleanup after 1 hour of inactivity
    • New obsidian://session-info resource for monitoring active sessions

  • Worker Thread Infrastructure: Foundation for parallel processing

    • Worker manager for handling CPU-intensive operations
    • Prepared infrastructure for offloading search and graph traversal
    • Non-blocking architecture to keep Obsidian UI responsive

  • Enhanced Connection Pool: Improved request handling

    • Queue-based processing with configurable limits
    • Session-aware request routing
    • Automatic resource cleanup and error recovery

Changed

  • Updated MCP server to support session headers (Mcp-Session-Id)
  • Enhanced debug logging to include session information
  • Improved request processing pipeline for better concurrency

Technical Details

  • Added ConnectionPool class for managing concurrent requests
  • Added SessionManager for tracking and expiring sessions
  • Added WorkerManager for future worker thread operations
  • Prepared semantic worker script for parallel processing

Previous Versions

See git history for changes before v0.5.8a