ktg-plugin-marketplace/plugins/llm-security/commands/mcp-inspect.md
Kjell Tore Guttormsen 03b8885b6e chore(llm-security): v7.7.2 — language consistency pass
~/.claude/CLAUDE.md specifies English for code and documentation,
Norwegian for dialog only. Norwegian had crept into surface text
across v7.5-v7.7. Translated to English in eight surfaces.

No scanner, hook, or behavior changes — purely surface text.

- 18 skill commands: the HTML Report-step now reads "HTML report:
  [Open in browser]" instead of "HTML-rapport: [Åpne i nettleser]"
- scripts/lib/report-renderers.mjs: key-stat labels, lede defaults,
  table headers, maturity-ladder descriptions, action-tier labels,
  clean buckets, dry-run/apply copy, and JS comments. Regex
  alternations /^high|^høy/ and /resolution|løsning/i preserved.
- playground/llm-security-playground.html: same renderer changes
  mirrored bit-identical, plus playground-only UI strings (catalog,
  breadcrumb aria-label, theme toggle, builder-modal hint,
  guide-panel "no projects yet", delete confirmation, alert/copy).
  Demo-state fixture content for dft-komplett-demo preserved
  (intentional Norwegian persona).
- agents/skill-scanner-agent.md + agents/mcp-scanner-agent.md:
  Generaliseringsgrense + Parallell Read-strategi sections translated
  to Generalization boundary + Parallel Read strategy.
- README.md: playground architecture prose + Recent versions table
  (v7.5.0 — v7.7.1).
- CLAUDE.md: v7.7.1 highlights translated, new v7.7.2 highlights
  added.
- ../../README.md: llm-security v7.5.0 — v7.7.1 bullets.
- ../../CLAUDE.md: llm-security catalog entry.
- docs/scanner-reference.md: six runnable-examples table cells.
- docs/version-history.md: new v7.7.2 entry. v7.5-v7.7 narrative
  sections left in original language (deferred per operator).
- Version bumped 7.7.1 → 7.7.2 in package.json,
  .claude-plugin/plugin.json, README badge + Recent versions,
  CLAUDE.md header + state, docs/version-history.md, playground
  renderHome hardcoded string, root README + CLAUDE.md llm-security
  entries.

Tests: 1820/1820 green. CLI smoke-test: 18/18 commandIds produce
>138 KB self-contained HTML. Browser-dogfood verified.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
2026-05-19 06:47:44 +02:00

2.7 KiB

name description allowed-tools model
security:mcp-inspect Connect to running MCP servers and scan live tool descriptions for injection, shadowing, and drift Bash, Read sonnet

/security mcp-inspect [path]

Live MCP runtime inspection — connects to each configured MCP stdio server via JSON-RPC 2.0, fetches tool/prompt/resource lists, and scans descriptions for injection patterns and tool shadowing.

Step 1: Run Scanner

Determine plugin root (parent of this commands/ folder) and target path from $ARGUMENTS (default .).

node <plugin-root>/scanners/mcp-live-inspect.mjs "<target>"

Parse stdout as JSON. If exit code is non-zero and no JSON output, show the error and STOP.

Step 2: Report

Display:

## MCP Live Inspection: <VERDICT>

**Servers:** <discovered> discovered | <contacted> contacted | <timed_out> timed out | <skipped_sse> SSE skipped | <failed> failed

### Server Details
| Server | Status | Tools | Prompts | Resources | Time |

For each finding (sorted critical → info):

### Findings (<count>)

| # | Severity | Title | OWASP | Recommendation |

If zero findings: "No injection, shadowing, or drift detected across N servers."

Step 3: Advisory

  • If servers_timed_out > 0: "N servers timed out (10s). Ensure they can start independently — check command, env vars, and args."
  • If servers_skipped_sse > 0: "N SSE servers skipped — live inspection requires stdio transport."
  • If servers_failed > 0: "N servers failed to start. Verify the command path exists and is executable."
  • If servers_discovered == 0: "No MCP servers found in config locations. Check .mcp.json, .claude/settings.json, or ~/.claude/settings.json."

Step 4: Combined Use

Mention: "For combined static + live analysis, use /security mcp-audit --live."

Step 5: HTML Report

After producing the markdown live-inspection report above:

  1. Compute a temp markdown path:

    node -p "require('path').join(require('os').tmpdir(), 'sec-mcp-inspect-' + Date.now() + '.md')"
    
  2. Use the Write tool to save the entire markdown report you just produced (Live Inspection banner + Server Details table + Findings table + Advisory) to that temp path. Verbatim.

  3. Run the renderer:

    node <plugin-root>/scripts/render-report.mjs mcp-inspect --in "<temp-md-path>"
    

    The CLI writes reports/mcp-inspect-<YYYYMMDD-HHmmss>.html relative to CWD and prints file:///abs/path.html on stdout.

  4. Append to your response (markdown link, no bare URL):

    HTML report: Open in browser

If the CLI exits non-zero, mention the error but do not block — the markdown report above is the primary deliverable.