files-stdio-mcp-server

Provides a Model Context Protocol (MCP) stdio server that exposes sandboxed, mount-scoped filesystem tools for an AI agent to explore directories, read files (with checksums), search by name/content, and safely edit files using checksum verification, unified diffs, and optional dry-run previews; also supports basic structural operations (rename/move/copy/mkdir/delete/stat) within configured roots.

Evaluated Mar 30, 2026 (0d ago)
Repo ↗ Infrastructure mcp filesystem sandbox agent-tools safe-editing stdio
⚙ Agent Friendliness
74
/ 100
Can an agent use this?
🔒 Security
56
/ 100
Is it safe for agents?
⚡ Reliability
28
/ 100
Does it work consistently?

Score Breakdown

⚙ Agent Friendliness

MCP Quality
85
Documentation
80
Error Messages
--
Auth Simplicity
95
Rate Limits
10

🔒 Security

TLS Enforcement
50
Auth Strength
20
Scope Granularity
90
Dep. Hygiene
60
Secret Handling
70

Security is primarily sandbox/mount based (FS_ROOTS/FS_ROOT), with strong safeguards for write safety (checksum verification, dry-run diff previews, and explicit guidance to re-read on checksum mismatch). No authentication/authorization model is described; if the MCP server is reachable by an untrusted agent, filesystem access would be governed only by the configured mounts. TLS is not applicable in stdio transport, so the score is moderate. Dependency hygiene is inferred from typical package usage; exact CVE status is not provided.

⚡ Reliability

Uptime/SLA
0
Version Stability
40
Breaking Changes
0
Error Recovery
70
AF Security Reliability

Best When

You want an MCP-enabled agent to work inside a narrow filesystem sandbox (e.g., a docs vault) with safety checks (checksum + dry-run + diffs).

Avoid When

You cannot constrain mounts/roots appropriately, or you require strong enterprise security/compliance features beyond local sandboxing.

Use Cases

  • Agent-assisted editing of text files (e.g., markdown, notes, documentation)
  • Browsing and searching within a knowledge base or vault without exposing arbitrary filesystem access
  • Safe, checksum-based updates with diff previews before applying changes
  • Structuring content: creating/moving/renaming files and folders under allowed roots

Not For

  • Non-text/binary file management (not the focus; includes MAX_FILE_SIZE and text-oriented read/write design)
  • High-integrity production file editing without human review (it helps, but is still an agent filesystem tool)
  • Access to arbitrary system paths outside configured FS_ROOTS/FS_ROOT

Interface

REST API
No
GraphQL
No
gRPC
No
MCP Server
Yes
SDK
No
Webhooks
No

Authentication

OAuth: No Scopes: No

No user authentication mechanism is described; access is controlled by local sandbox configuration (FS_ROOTS/FS_ROOT). Security relies on mount restrictions rather than auth/identity.

Pricing

Free tier: No
Requires CC: No

Self-hosted open-source style package; pricing not described.

Agent Metadata

Pagination
none
Idempotent
False
Retry Guidance
Not documented

Known Gotchas

  • Must use relative paths within mounts; absolute paths should be rejected by the sandbox.
  • To apply edits safely, agents should read first and use fs_write with the latest checksum; checksum mismatch requires re-reading.
  • Deletion is limited (single file or empty directory only); recursive delete is intentionally not supported for safety.
  • Dry-run should be used to preview diffs for destructive/meaningful changes.

Alternatives

Full Evaluation Report

Comprehensive deep-dive: security analysis, reliability audit, agent experience review, cost modeling, competitive positioning, and improvement roadmap for files-stdio-mcp-server.

AI-powered analysis · PDF + markdown · Delivered within 30 minutes

$99

Package Brief

Quick verdict, integration guide, cost projections, gotchas with workarounds, and alternatives comparison.

Delivered within 10 minutes

$3

Score Monitoring

Get alerted when this package's AF, security, or reliability scores change significantly. Stay ahead of regressions.

Continuous monitoring

$3/mo

Scores are editorial opinions as of 2026-03-30.

6435
Packages Evaluated
19968
Need Evaluation
586
Need Re-evaluation
Community Powered