A Map, Not a Megadump: Sentinel-IR-MCP 0.2.0 Adds TypeScript, 0.2.1 Makes Missing Analysis Visible

Description and use

Sentinel-IR-MCP 0.2.1 exposes incomplete and unreadable files in the directory summary, enhancing code navigation and assessment for AI agents.

Problem

AI agents reading entire codebases consume excessive context and time; missing or unreadable files can hide uncertainty in analysis reports.

Outcome

Sentinel-IR-MCP 0.2.1 provides transparent reporting of incomplete and unreadable files, improving AI agents’ ability to navigate and assess codebases safely.

Workflow

• 1. Start Sentinel-IR-MCP server with Node.js 20+ using npx and project path.
• 2. Configure MCP client (e.g., Claude Desktop or Cursor) with server command and arguments.
• 3. Use MCP tools: analyze_code for source text, analyze_file for a single file, analyze_directory for bulk analysis.
• 4. Inspect the returned JSON: check files list, incomplete list, statusCounts, and skipped.unreadable for completeness and safety assessment.

Evidence

Version 0.2.1 made a different, less flashy but equally important improvement: it made incomplete and skipped files visible in the directory summary, rather than letting a low-risk-looking entry hide uncertainty or one unreadable file ruin the whole result.

Source

The new fields are: "incomplete": paths whose analysis status is not "complete", "statusCounts": totals for "complete" and "incomplete" among analyzed results, "skipped.unreadable": paths that could not produce a normal per-file result because the read/analysis operation failed.

Source