Vestige
Analyzes Electron applications via the JavaScript provider, building an AST-based module graph and inventorying Electron IPC channels exposed by the app.
Inspects JavaScript/TypeScript sources to produce an Express-style route inventory, enumerating HTTP endpoints declared in the application.
Analyzes JavaScript/TypeScript sources and directories in-process (no external tools), producing an AST module graph of imports, exports, requires, and dynamic imports, plus manifest and dependency facts, with evidence-linked findings.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Vestige analyze the registered Electron app and report its IPC channels"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Vestige
Evidence-driven reverse engineering for AI agents.
An open, independent MCP platform that lets compatible AI coding agents inspect authorized software artifacts, investigate their structure, and produce verifiable, evidence-linked technical reports.
Why Vestige
Most analysis tools answer questions. Vestige answers questions with receipts: every claim
an agent makes about a binary or application is backed by a content-addressed evidence record
that is re-authenticated whenever it is read. Accuracy is treated as a measurable engineering
objective — benchmarks run against deterministic fixtures with embedded ground truth, and the
results ship in docs/BENCHMARKS.md only after actually running.
Vestige is an independent project — inspired by the engineering patterns of
morluto/rea (MIT), but with its own architecture,
implementation, identity, and test suite. See
docs/ASSESSMENT-REA.md for the full reference assessment and the
improvement matrix that motivated this design.
Related MCP server: REA
Quick start
# Requires Node.js >= 20.19 (tested on 24.x)
npm install
npm run build
npm test # 137+ tests: unit, integration, security, regression, e2e
# Register a target and analyze it via the CLI
node dist/cli/main.js register ./path/to/app --label "sample app"
node dist/cli/main.js providers
node dist/cli/main.js analyze <artifactId> --provider javascript
node dist/cli/main.js findings <artifactId>
node dist/cli/main.js doctorMCP client configuration
stdio (recommended for local agents):
{
"mcpServers": {
"vestige": {
"command": "node",
"args": ["/absolute/path/to/vestige/dist/cli/main.js", "mcp"],
"env": {
"VESTIGE_DATA_DIR": "/absolute/path/to/vestige-data",
"VESTIGE_LOG_LEVEL": "info"
}
}
}
}Streamable HTTP (opt-in, remote deployments):
VESTIGE_HTTP_TOKEN=$(openssl rand -hex 32) \
VESTIGE_HTTP_HOST=0.0.0.0 VESTIGE_HTTP_PORT=8420 \
node dist/cli/main.js mcp --httpHTTP requires a bearer token for any non-loopback bind (constant-time compared), validates
Host headers against an allowlist (DNS-rebinding defense), and rejects disallowed Origins.
Clients connect to http://host:8420/mcp with Authorization: Bearer <token>.
What it can analyze (today, with tests)
Provider | Formats | Capabilities | External tool? |
| JS/TS sources, directories, Electron apps | AST module graph (imports/exports/requires/dynamic), Electron IPC channels, express-style route inventory, manifest/dependency facts | No (in-process acorn) |
| ELF (32/64), PE (PE32/PE32+), Mach-O (incl. FAT) | Headers, sections, symbols/functions, PE imports & exports, Mach-O dylibs, strings with offsets | No (pure TypeScript) |
| ELF, PE, Mach-O | Headless decompilation, function inventory, call graph | Yes — Ghidra 11.x + Java 21+ (optional; availability-checked with install guidance) |
Unsupported formats are rejected honestly: Vestige never advertises a capability it cannot
demonstrate. Provider availability is always queryable (vestige_list_providers,
vestige doctor) and includes actionable installation guidance.
Evidence model
Every significant finding carries:
Finding ID (
vf_…) and Evidence ID (ve_…) — content-derived, tamper-checked on readTarget artifact ID + SHA-256 commitment taken at registration
Provider name/version and external tool version when applicable
Method used to obtain the observation, timestamp, and typed location (file offset / virtual address / artifact path / section)
Confidence class with explicit semantics:
observed-fact·static-inference·runtime-observation·ai-interpretation·hypothesisLimitations declared up front by the provider
Evidence is stored content-addressed; re-analysis of identical content is served from a verified cache. Reports (markdown/JSON) are deterministic and reproducible.
Safety by default
Explicit registration required — nothing is analyzed unless the operator registered it
Read-only analysis — original artifacts are never modified; generated files live in an isolated workspace
Canonical path validation —
realpathresolution, symlink policy, allowed-roots enforcement, traversal rejectionNever executes unknown targets — runtime observation is not implemented in v0.1 by design; static analysis is preferred
Shell-free process execution — argv arrays only, output caps, hard timeouts, SIGTERM→SIGKILL stop ladder, abort support
Central redaction — logs and captured output pass a secret-scrubbing layer before persistence
See docs/THREAT-MODEL.md and SECURITY.md.
Architecture (60-second tour)
graph TD
CLI["CLI (src/cli)"] --> PLAT["Composition root (src/server/context.ts)"]
MCP["MCP server (src/server): stdio + streamable HTTP"] --> PLAT
subgraph services["Application services (src/services)"]
AN["AnalysisService<br/>jobs · cache · routing"]
TR["TargetRegistry<br/>admission + SHA-256 commitment"]
ER["EvidenceStore<br/>content-addressed objects"]
PR["ProviderRegistry<br/>capabilities + availability"]
end
PLAT --> services
AN --> PROV
subgraph PROV["Providers (src/providers) — depend on interfaces only"]
JS["javascript<br/>acorn AST"]
NAT["native<br/>ELF · PE · Mach-O"]
GHI["ghidra<br/>headless bridge"]
end
PROV --> SEC["security: PathGuard · Workspace"]
PROV --> PROC["process: runBounded (shell-free, bounded)"]
PROV --> DOM["domain: Zod contracts · evidence · findings"]Full details, module boundaries, and dependency rules: docs/ARCHITECTURE.md.
Documentation
Document | Contents |
Modules, dependency direction, data flow | |
Tool/resource/prompt reference with schemas | |
Provider guides, Ghidra setup, troubleshooting | |
Assets, attackers, mitigations, residual risks | |
Measured accuracy & latency (from actual runs) | |
Tested vs intended platform support | |
Reference repository assessment & improvement matrix | |
Security policy and responsible use | |
Version history | |
Development workflow and quality gates |
Verification status (this checkout)
The following were actually executed in the development environment (Linux x64, Node 24.22.1):
Gate | Command | Result |
Type check |
| ✅ pass (strict) |
Build |
| ✅ pass |
Tests |
| ✅ 137 passed / 0 failed (unit 62, integration 17, security 12, regression 5, e2e 1, fixtures & parsers 40) |
Benchmarks |
| ✅ recorded in |
Platform notes: Windows and macOS are intended targets with platform-neutral code paths, but
only Linux was exercised in this checkout — see docs/COMPATIBILITY.md.
License
MIT — see LICENSE. Third-party notices: THIRD_PARTY_NOTICES.md. Vestige is not affiliated with, endorsed by, or an official release of the REA project.
Available Tools
11 toolsvestige_analyzeStart analysis jobARead-only
Start an asynchronous analysis job on a registered target. Returns a jobId immediately; poll vestige_job_status. Progress notifications are sent when the client supplies a progressToken in _meta.
Inputs: artifactId (required), provider, capabilities.
vestige_diagnosticsRun self-diagnosticsARead-only
Report versions, configuration facts, data directories, and provider availability. Never contains secrets.
Inputs: no named parameters.
vestige_export_reportExport analysis reportBRead-only
Export a deterministic markdown or JSON report for a target into the Vestige workspace and return its absolute path.
Inputs: artifactId (required), format.
vestige_get_evidenceGet evidence recordARead-only
Fetch one evidence record by id. Records are re-authenticated against their content digest before being served.
Inputs: evidenceId (required).
vestige_get_targetGet target metadataCRead-only
Get one registered target's metadata plus which providers support it.
Inputs: artifactId (required).
vestige_job_cancelCancel analysis jobBRead-only
Cancel a queued or running analysis job. Evidence already persisted is retained and the job becomes 'cancelled'.
Inputs: jobId (required).
vestige_job_statusGet job statusBRead-only
Get the status, progress, and result counts of an analysis job.
Inputs: jobId (required).
vestige_list_findingsList findingsBRead-only
List evidence-linked findings for a target, optionally filtered by kind or confidence, with pagination.
Inputs: artifactId (required), kind, limit, offset, +1 more.
vestige_list_providersList analysis providersARead-only
List providers with capabilities and live availability (including installation guidance when a tool like Ghidra is missing).
Inputs: no named parameters.
vestige_list_targetsList registered targetsARead-only
List all registered analysis targets with identity digests.
Inputs: no named parameters.
vestige_register_targetRegister analysis targetARead-only
Register a file or directory for analysis. Validates the path (canonicalization, symlink policy, allowed roots), computes SHA-256, and identifies the format. Read-only: the artifact is never modified.
Inputs: path (required), label.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 11 tools
Each tool targets a distinct resource or action: target registration/listing/retrieval, provider discovery, asynchronous analysis, job status/cancellation, findings listing, evidence retrieval, report export, and diagnostics. Boundaries are clear, especially between starting an analysis and checking its job status.
Most tools use the consistent vestige_ prefix with verb_noun naming (register_target, list_targets, get_target, list_providers, list_findings, get_evidence, export_report). Minor deviations appear in vestige_analyze, vestige_diagnostics, and noun-first job operations (vestige_job_status, vestige_job_cancel), but the set remains predictable.
Eleven tools is well-scoped for an artifact analysis workflow, covering registration, provider discovery, asynchronous job control, findings/evidence access, reporting, and diagnostics without obvious padding.
The surface covers the core lifecycle from target registration through analysis, job monitoring, evidence/finding inspection, and report export. Minor gaps exist around target lifecycle management such as unregistering/removing registered targets, but agents can work around this.
Maintenance
Related MCP Connectors
Browser-local and CLI static evidence for deployed AI model artifacts.
Governed security assessments, findings and signed reports for your domains, from your AI agent.
1Code intelligence platform for AI agents. 20 tools for architecture, security & impact analysis.
CVE lookups (NVD) and dependency-manifest audits (OSV) for AI agents. No API keys.
Related MCP Servers
- AlicenseNot gradedqualityBmaintenanceEnables AI clients to inspect local software structure, code, and runtime behavior through evidence-backed analysis, including sources, coverage, and limitations.1MIT
- AlicenseNot gradedqualityNot gradedmaintenanceEnables AI agents to inspect and reverse engineer local software targets, including binaries, JavaScript/Electron apps, .NET assemblies, and websites, returning evidence-backed findings for understanding and building similar features.377,174 npmMIT
- AlicenseNot gradedqualityBmaintenanceEnables MCP-compatible AI assistants to inspect local application internals by analyzing JavaScript/Electron/ASAR and .NET artifacts, tracing dependencies, interfaces, and data flows, and preserving traceable evidence.14MIT
- AlicenseNot gradedqualityBmaintenanceConnects AI agents to local analysis tools that inspect native binaries, JavaScript/Electron apps, .NET assemblies, and websites without source code, so features can be explained, traced, and rebuilt. Every finding is returned with its supporting evidence and stated limitations.1MIT