4da MCP Server

Install

Works with Claude Code, Codex, VS Code, Cursor, Windsurf, Claude Desktop. Read the repository before you give any tool access to your data.

npx agentmarket install connector/4da-systems-4da-mcp-server

About

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this server can do and what it asks for, from its tool list, install config and code. Checked Oct 26, 2026.

Tools

8 read · 4 write · 0 delete · 4 not yet classified

Secrets

None asked for

Install scripts

None found

Runs as

Node.js package

Tools (16)

Read from the source code; the running server may differ. Descriptions are the server’s own words; hidden characters are shown as [U+…].

• agent_memory

  Cross-agent persistent memory. Actions: store (save a memory), recall (search by subject), recall_by_tags (search by tags), get_recent (memories since timestamp). What one agent learns, all agents can access.

• check_decision_alignment

  Check if a technology or pattern aligns with the developer's active decisions. Call BEFORE suggesting major tech changes. Returns alignment status, relevant decisions, and any conflicts.

• decision_memory

  Manage developer decisions. Actions: record (create new decision), list (query decisions), check_alignment (check if tech/pattern conflicts with decisions), update (modify a decision), supersede (replace old decision with new one).

• dependency_check

  Is this dependency version safe to write? Call BEFORE you write any version into a manifest or lockfile: adding, pinning, upgrading, bumping or downgrading a package, or choosing a version (npm, crates.io, PyPI, Go). Pass the exact version you intend to write, or the constraint you were given ("7.x", "^1.20", "~=3.1", "v1") to get the smallest and newest safe release in it. Write only a version that came back `proceed`. Your memory of which release is safe is out of date; this reads the registry and OSV.dev now. Per item: `verdict` (proceed | wait | review | avoid | unknown), a one-line `reas…

• dependency_health

  Assess health of project dependencies — version freshness, deprecation status, known CVEs. Auto-detects stack from lock files. Covers npm, Rust, Python, and Go. For npm, `installedVersion` appears beside `currentVersion` when node_modules holds a different version than the lockfile. Two populations: `totalDeps`, `outdatedCount` and `deprecatedCount` are over the DIRECT dependencies (registry rows; devDependencies only with include_dev); `vulnerableCount` is distinct vulnerable packages in the full installed tree of `treeDeps` (the same set as vulnerability_scan), `vulnerableDirectCount` the di…

• developer_dna

  Export your Developer DNA — a comprehensive profile of your tech identity built from project scans, explicit interests, and local interaction history. Includes primary stack, adjacent tech, top dependencies, engaged topics, blind spots, source engagement rates, and aggregate stats.

• ecosystem_pulse

  Live ecosystem news relevant to your tech stack. Surfaces trending Hacker News discussions filtered by your detected technologies. Updated on server startup.

• get_actionable_signals

  Get actionable signals classified from recent content. Categorizes items into signal types: security_alert, breaking_change, tool_discovery, tech_trend, learning, competitive_intel. Each signal has a priority level (critical/high/medium/low) based on signal type, relevance score, and tech stack match. Use this to get prioritized, actionable intelligence from 4DA's feed.

• get_context

  Get the user's context - what 4DA knows about them. Returns information about: - Static Identity: User-declared role, tech stack, domains, interests, and exclusions - ACE Context (optional): Auto-detected technologies and active topics from recent activity - Learned Preferences (optional): retained compatibility field; implicit preference learning no longer contributes to scoring Use this to understand what the user is working on and interested in.

• get_relevant_content

  Get filtered relevant content from 4DA's personalized feed. Returns content items that match the user's explicit interests, tech stack, and ACE-detected context. Items are scored based on: - Explicit interests declared by the user - Tech stack and domains - ACE-detected context (recent files, projects, git activity) Each item includes necessity fields (necessity_score, necessity_reason, necessity_category, necessity_urgency) indicating how critical the item is — e.g., security vulnerabilities affecting your deps, breaking changes in your stack, or deprecation notices. These are populated fro…

• knowledge_gaps

  Detect knowledge gaps by cross-referencing your project dependencies with source items you haven't engaged with. Identifies things you should know about but might have missed.

• record_feedback

  Record explicit user interaction history for a content item. Feedback actions: - "click": User clicked/opened the item - "save": User saved/bookmarked the item, or said it is relevant (also records a relevance label) - "dismiss": User dismissed the item - "mark_irrelevant": User said the item is not relevant (also records a relevance label) Only "save" and "mark_irrelevant" are relevance labels. Use them only when the user actually says whether the item matters to them. This does not train content preferences. Recorded in the 4DA desktop app's own interactions/feedback tables.

• upgrade_impact

  What changes if you upgrade ONE dependency from its installed version to a target — call before bumping a package, especially across a major. Lists the releases in between (publish dates, npm deprecations, crates yanks), reads the changelog shipped inside the target version's registry archive, returns its entries with the heading each sits under (`under`), flags likely breaking / deprecation / security entries from the changelog's headings and wording, and marks the ones that mention symbols or route syntax THIS project uses (touches_your_code). The flags are a pre-sort, not a verdict: read ev…

• upgrade_planner

  Ranked dependency upgrade plan for one project — call before upgrading, adding, or auditing any dependency to pick the safest order. `project_path` picks the project (default: the one the server was started in; "*" for every project). When the 4DA desktop app has computed its plan, returns that plan's work order (provenance app_plan): per package the ecosystem, each installed version with its minimum clean target, upgrade type (patch/minor/major), the projects that hold it (direct/dev) and the mechanism (manifest_bump, lockfile_or_parent_update, mixed, no_fix); flagged stale past its freshness…

• vulnerability_scan

  Scan project dependencies for known vulnerabilities (CVEs) using OSV.dev. Returns severity, fix versions, and upgrade recommendations for npm, Rust, Python, Go, Ruby (Gemfile.lock) and PHP (composer.lock). Maven, Gradle, .NET and other ecosystems are not read: their manifests are listed under `coverage.not_covered`, never silently skipped; test-fixture lockfiles are counted under `coverage.not_scanned`. Zero config — scans the complete resolvable lockfile package set, including transitives, and for npm also the version node_modules actually holds when it differs from the lockfile (install_drif…

Signals

Stars

0

Forks

0

Open issues

1

Weekly downloads

156

Last commit

Oct 15, 2026 (15 days ago)

Last release

Oct 12, 2026 (18 days ago)

Last package publish

Oct 12, 2026 (18 days ago)

From GitHub (fetched Oct 24, 2026) and the package registry.

Details

Registry name

io.github.4DA-Systems/4da-mcp-server

Version

6.2.4

Auth

No auth

Packages

npm @4da/mcp-server

License

Apache-2.0

Repository

github.com/4DA-Systems/4da-mcp-server

First seen

Oct 24, 2026

Last checked

Oct 25, 2026

Source

Official MCP Registrygithub-graphql