Atlas / MCP servers / kirachon / Context Engine

Context EngineBLOCK

Name: Context Engine
Author: kirachon

mcp/kirachon/context-engine-2

MCP server by kirachon · ai assistant · code review · context engine · Audited 2026-10-08

⚑ Report / re-scan Source ↗

Local-native Context Engine MCP server for codebase indexing, semantic retrieval, planning, and review workflows.

Verdict

BLOCK

Grade

Trust score

57 /100

Exposed tools

47 37r · 7w · 3d

Transport

stdio · streamable-http

License

—

Stars

Overview

From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.

A local-first, agent-agnostic Model Context Protocol (MCP) server for workspace indexing, retrieval, planning, and review workflows, with a setup path that works well for Codex and other OpenAI-powered agents.

New here? Start with the beginner quick start below. If you want client-specific setup help, see docs/MCPCLIENTSETUP.md. If you are on Windows, see docs/WINDOWSDEPLOYMENTGUIDE.md. Historical docs live in docs/archive/INDEX.md if you need the old planning and migration notes.

OpenAI / Codex Showcase

If you want to see what this project demonstrates for OpenAI-style agent workflows, start here:

Local workspace indexing and retrieval
Review and planning workflows layered on top of the same MCP server
Beginner-friendly install and client setup
Windows support and copy-paste setup examples
AI-agent-friendly instructions for self-setup

Why This Matters

It shows how an OpenAI-powered agent can connect to a real workspace and start using tools right away.
It combines retrieval, review, and planning in one MCP server instead of relying on one-off scripts.
It gives both humans and AI agents a simple, repeatable setup path, which makes demos and onboarding easier.

Fastest demo path:

npm install
npm run build
codex mcp add context-engine -- node dist/index.js

Then in Codex, confirm the tools are visible and try:

use semantic_search to find authentication logic

Beginner Quick Start

If you just want to get Context Engine running locally, follow these steps:

Install Node.js 18+.
Clone this repository and open it in a terminal at the repo root.
Install dependencies:

npm install

Build the server:

npm run build

Run the verification checks:

npm run verify

Start the MCP server:

node dist/index.js

By default,

Read from source at commit a6b1576c8af7OBSERVED · 2026-10-08

Connect

Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control. Replace the environment placeholders with a token scoped to the least it needs.

claude-code

claude mcp add context-engine --env CONTEXT_ENGINE_HTTP_AUTH_ENABLED=${CONTEXT_ENGINE_HTTP_AUTH_ENABLED} --env CONTEXT_ENGINE_HTTP_AUTH_TOKENS=${CONTEXT_ENGINE_HTTP_AUTH_TOKENS} --env REACTIVE_TOKEN_BUDGET=${REACTIVE_TOKEN_BUDGET} -- npx -y [email protected]

claude-desktop

{
  "mcpServers": {
    "context-engine": {
      "command": "npx",
      "args": [
        "-y",
        "[email protected]"
      ],
      "env": {
        "CONTEXT_ENGINE_HTTP_AUTH_ENABLED": "${CONTEXT_ENGINE_HTTP_AUTH_ENABLED}",
        "CONTEXT_ENGINE_HTTP_AUTH_TOKENS": "${CONTEXT_ENGINE_HTTP_AUTH_TOKENS}",
        "REACTIVE_TOKEN_BUDGET": "${REACTIVE_TOKEN_BUDGET}"
      }
    }
  }
}

Exposed tools (47)

37 read · 7 write · 3 destructive. Blast radius: 3 tools can delete or overwrite — an agent that can be talked into calling a tool can be talked into calling this one.

Tool	Risk	Description
`Coverage`	read	Expand test contracts
`add_memory`	write	Store a memory for future sessions. Memories are persisted as markdown files and automatically retrieved via semantic search when relevant. Categories: - \
`categories`	read	Optional comma-separated category list.
`check_invariants`	write	Run YAML invariants deterministically against a unified diff (no LLM).
`clarifications`	read	Optional JSON object string of clarification answers.
`clear_index`	destructive	Remove saved index state and clear caches without rebuilding.
`compare_plan_versions`	read	Generate a diff between two versions of a plan.
`complete_step`	read	Mark a step as completed with optional notes.
`create-plan`	write	Build a planning request using the repo planning templates.
`current_plan`	read	Existing plan JSON string.
`custom_instructions`	read	Optional extra review instructions.
`delete_plan`	destructive	Delete a saved plan from storage.
`diff`	read	Unified diff to review.
`enhance-request`	read	Build a request-enhancement prompt using the repo prompt-enhancement templates.
`fail_step`	read	Mark a step as failed with error details.
`feedback`	read	Optional refinement feedback text.
`find_callees`	read	Return deterministic callees of a known function or method symbol. This tool prefers persisted graph call edges and reports explicit degraded-mode receipts when it had to fall back to heuristic extraction.
`find_callers`	read	Return deterministic callers of a known function or method symbol. This tool prefers persisted graph call edges and falls back explicitly when graph coverage is unavailable or incomplete. Use when you want call sites for one symbol without the broader combined output of call_relationships.
`get_review_status`	read	Get the current status and progress of a reactive review session. Returns: - Session status (active, paused, completed, cancelled, error) - Progress percentage and step counts - Findings count - Telemetry data (elapsed time, tokens used, cache hit rate)
`get_review_telemetry`	read	Get detailed telemetry data for a review session. Returns: - Token usage statistics - Cache hit/miss rates - Execution timing per step - Reactive configuration in use
`impact_analysis`	read	Estimate the direct change surface of a known symbol using graph-backed definition, reference, and call-edge data. This v1 analysis is intentionally bounded to direct references and call edges so degraded-mode behavior stays explicit and deterministic.
`index_status`	read	Retrieve current index health metadata (status, last indexed time, file count, staleness).
`list_memories`	read	List all stored memories, optionally filtered by category. Shows file stats, entry counts, and content preview for each memory category.
`list_plans`	read	List all saved plans with optional filtering.
`load_plan`	read	Load a previously saved plan by ID or name.
`pause_review`	read	Pause a running reactive review session. The review can be resumed later with resume_review. Useful for: - Freeing up resources temporarily - Allowing manual intervention - Stopping execution before a problematic step
`refine-plan`	read	Build a plan-refinement request using the repo refinement templates.
`reindex_workspace`	destructive	Clear current index state and rebuild it from scratch.
`request_approval`	write	Create an approval request for a plan or specific steps.
`respond_approval`	read	Respond to a pending approval request (approve, reject, or request modifications).
`resume_review`	read	Resume a paused reactive review session. Continues execution from where it was paused.
`review-diff`	read	Build a code-review request for a diff using the repo review templates.
`review_auto`	read	Smart wrapper that chooses review_diff when a diff is provided; otherwise chooses review_git_diff for the current git workspace.
`review_diff`	read	Enterprise-grade diff-first review with deterministic preflight and structured JSON output.
`review_git_diff`	read	Review code changes from git automatically. This tool combines git diff retrieval with AI-powered code review. It automatically: 1. Retrieves the diff from git based on the target 2. Analyzes the changes for issues 3. Returns structured findings with confidence scores Target Options: -
`review_memory_suggestions`	read	Review draft memory batches and perform idempotent batch actions. Batch listing enforces the Phase 1 cap (target 3-5, hard cap 5). Approved suggestions remain review-state only in this phase; durable writes stay in the existing memory tools. Undo is pre-promotion only in Phase 1.
`rollback_plan`	read	Rollback a plan to a previous version.
`run_static_analysis`	write	Run local static analyzers (tsc and optional semgrep) and return structured findings.
`save_plan`	write	Save a plan to persistent storage for later retrieval and execution tracking.
`start_step`	write	Mark a step as in-progress to begin execution.
`symbol_references`	read	Find non-declaration usages of a known identifier across the local codebase. Use this tool when you need to: - Locate call sites or consumers of a known function, class, or constant - Exclude declaration hits from identifier-style navigation - Narrow usage lookup with include_paths or exclude_paths
`tool_manifest`	read	Discover available tools and capabilities exposed by the server.
`trace_symbol`	read	Trace a known symbol across its canonical definition, non-declaration references, and direct call edges. This tool composes the graph-backed symbol navigation surfaces into one deterministic response and makes degraded-mode behavior explicit per stage.
`view_history`	read	View version history for a plan.
`view_progress`	read	View execution progress for a plan.
`visualize_plan`	read	Generate diagrams from an implementation plan. Use this to visualize the plan
`why_this_context`	read	Explain why files were selected into a context bundle using the shared retrieval provenance and explainability contract. This tool reuses the same selection vocabulary exposed by graph-aware retrieval and get_context_for_prompt instead of inventing a parallel explanation path.

Trust audit

BLOCKgrade F · trust 57/100 Do not install this without reading the findings. The audit found something that could harm you or your machine.

Layer	What it checks	Result
L0	Provenance & inventory	WARN
L1	Static analysis of the code	FAIL
L2	Instruction surface (what it tells the agent)	FAIL
L3	Class-specific surface	WARN
L4	Behavioural (sandbox)	SKIPPED

What the source does

Filesystem: declared (5 observation(s))
Network: declared (3 observation(s))
Shell: none-observed
Dependencies: not all pinned
Secrets in source: found

Findings (25)

HIGHNetwork egress · net.metadata_ip · CWE-200, CWE-319

src/mcp/tooling/externalGrounding.ts:157

normalized === '169.254.169.254' ||

Why it matters. cloud metadata endpoint: the classic SSRF credential grab

HIGHNetwork egress · net.metadata_ip · CWE-200, CWE-319

src/mcp/tooling/externalGrounding.ts:158

normalized === 'metadata.google.internal'

Why it matters. cloud metadata endpoint: the classic SSRF credential grab

HIGHPrompt injection · prompt.override · CWE-94, CWE-1427

docs/VIBE_CODER_MEMORY_REVIEW_CONTRACT.md:92

- Bulk save must not bypass per-item safety rules.

Why it matters. asks the agent to drop prior instructions or safety

Fix. remove the instruction

MEDIUMInventory / provenance · inv.binary · CWE-1104

context-engine.zip

context-engine.zip

Why it matters. a compiled or binary member cannot be reviewed from source

Fix. ship source, or explain the binary in the README

MEDIUMHard-coded secrets · secret.generic · CWE-798, CWE-321

docs/archive/EXAMPLES.md:994

"code_snippet": "const secret = 'my-super-secret-key-12345';"

MEDIUMHard-coded secrets · secret.generic · CWE-798, CWE-321

tests/tools/reactiveReview.test.ts:541

content: 'const apiKey = "sk-proj-abc123def456ghi789jklmno";',

MEDIUMFilesystem / path · mcp.destructive_tools · CWE-22, CWE-59

clear_index, delete_plan, reindex_workspace

Why it matters. 3 tool(s) can delete or overwrite

Fix. prefer a read-only mode or scoped tokens; the page states the blast radius

LOWInventory / provenance · inv.binary · CWE-1104

examples/skills/context-engine-mcp.skill

context-engine-mcp.skill

Why it matters. a compiled or binary member cannot be reviewed from source

Fix. ship source, or explain the binary in the README

LOWInventory / provenance · inv.binary · CWE-1104

examples/skills/flutter-playwright-e2e.skill

flutter-playwright-e2e.skill

Why it matters. a compiled or binary member cannot be reviewed from source

Fix. ship source, or explain the binary in the README

LOWInventory / provenance · inv.hidden_file · CWE-1104

.contextignore

.contextignore

Why it matters. hidden member outside the usual dotfiles

Fix. review its purpose

LOWInventory / provenance · inv.hidden_file · CWE-1104

.review-invariants.yml

.review-invariants.yml

Why it matters. hidden member outside the usual dotfiles

Fix. review its purpose

LOWInventory / provenance · inv.hidden_file · CWE-1104

vscode-extension/.vscodeignore

.vscodeignore

Why it matters. hidden member outside the usual dotfiles

Fix. review its purpose

LOWInventory / provenance · inv.no_license · CWE-1104

Why it matters. no LICENSE file and no repo licence

Fix. add a licence