Persistent memory and project context for AI coding agents — context graph, semantic search, decision log, roadmap, and cross-tool continuity via local MCP server
Config is the same across clients — only the file and path differ.
{
"mcpServers": {
"codevira": {
"cwd": "/path/to/your-project",
"args": [],
"command": "codevira"
}
}
}Are you the author?
Add this badge to your README to show your security score and help users find safe servers.
Persistent memory and project context for AI coding agents — context graph, semantic search, decision log, roadmap, and cross-tool continuity via local MCP server
This server supports HTTP transport. Be the first to test it — help the community know if it works.
Five weighted categories — click any category to see the underlying evidence.
pip Vulnerable to Inclusion of Functionality from Untrusted Control Sphere
pip prior to version 26.1 would run self-update check functionality after installing wheel files which required importing well-known Python modules names. These module imports were intentionally deferred to increase startup time of the pip CLI. The patch changes self-update functionality to run before wheels are installed to prevent newly-installed modules from being imported shortly after the installation of a wheel package. Users should still review package contents prior to installation.
pip has an interpretation conflict due to handling both concatenated tar and ZIP files as ZIP files
pip handles concatenated tar and ZIP files as ZIP files regardless of filename or whether a file is both a tar and ZIP file. This behavior could result in confusing installation behavior, such as installing "incorrect" files according to the filename of the archive. New behavior only proceeds with installation if the file identifies uniquely as a ZIP or tar archive, not as both.
>= 0source →pip Path Traversal vulnerability
When pip is installing and extracting a maliciously crafted wheel archive, files may be extracted outside the installation directory. The path traversal is limited to prefixes of the installation directory, thus isn't able to inject or overwrite executable files in typical situations.
pip's fallback tar extraction doesn't check symbolic links point to extraction directory
When extracting a tar archive pip may not check symbolic links point into the extraction directory if the tarfile module doesn't implement PEP 706. Note that upgrading pip to a "fixed" version for this vulnerability doesn't fix all known vulnerabilities that are remediated by using a Python version that implements PEP 706. Note that this is a vulnerability in pip's fallback implementation of tar extraction for Python versions that don't implement PEP 706 and therefore are not secure to all vulne
PYSEC-2023-228
When installing a package from a Mercurial VCS URL (ie "pip install hg+...") with pip prior to v23.3, the specified Mercurial revision could be used to inject arbitrary configuration options to the "hg clone" call (ie "--config"). Controlling the Mercurial configuration can modify how and which repository is installed. This vulnerability does not affect users who aren't installing from Mercurial.
Click any tool to inspect its schema.
decision_replayBrowse decision timeline as terminal / markdown / HTML; available as MCP resource for Claude Desktop
codevira://decisions
This server is missing a description.If you've used it, help the community.
Add informationBe the first to review
Have you used this server?
Share your experience — it helps other developers decide.
Sign in to write a review.
Others in search
Web and local search using Brave Search API
Production ready MCP server with real-time search, extract, map & crawl.
Multi-engine MCP server, CLI, and local daemon for agent web search and content retrieval — skill-guided workflows, no API keys.
MCP server for Firecrawl — search, scrape, and interact with the web. Supports both cloud and self-hosted instances. Features include web search, scraping, page interaction, batch processing, and LLM-powered content analysis.
MCP Security Weekly
Get CVE alerts and security updates for Codevira and similar servers.
Start a conversation
Ask a question, share a tip, or report an issue.
Sign in to join the discussion.
One memory layer for every AI coding tool you use. Switch between Claude Code, Cursor, Windsurf, and Antigravity without losing context, decisions, or progress.
Built for solo developers working on local projects with AI agents. Codevira gives every AI tool you use access to the same persistent project memory — so you stop re-explaining your codebase every session, stop losing carefully-made decisions, and stop burning tokens on re-discovery.
Works with: Claude Code · Claude Desktop · Cursor · Windsurf · Google Antigravity · OpenAI Codex · GitHub Copilot · Continue.dev · Aider · any MCP-compatible AI tool
Edit, Write, prompt submit, session start). Decisions you mark do_not_revert block silent regressions. Out-of-scope edits get warned. The AI literally cannot undo your protected choices without surfacing the conflict to you first.pipx install codevira && codevira setup — auto-detects every AI tool you have and configures all of them. No JSON to hand-edit, no per-IDE script, no team server to spin up.~/.codevira/ on your machine. No cloud, no account, no telemetry, no SaaS.codevira insights shows which past decisions held up vs got reverted across your real git history. codevira budget reports per-session AI token spend. You can audit what the memory layer is actually costing and earning.If you code with AI agents on a project longer than a week, you've felt all of these:
Every new chat starts from zero. The AI doesn't know your architecture, your conventions, your "we don't do it that way" decisions. You waste the first 10 minutes (and thousands of tokens) catching it up — only to do it again tomorrow.
Last week you debugged a tricky retry policy for 3 hours. Today's AI session refactors it to a simpler version because it has no idea why the complexity exists. Now it's broken again.
You started planning in Claude Code. Switched to Cursor for autocomplete. Opened Antigravity to run tests. Three different agents, three different blind copies of your project state. Nothing carries over.
Your AI agent reads the same 12 files every session before doing any actual work. You're paying API costs for the same lookups, over and over.
Codevira is a persistent memory layer that fixes all four — for every AI tool, on every project, on your local machine.
v2.1.1 shipped hybrid search (BM25 + semantic) but without a similarity floor, so off-topic prompts kept surfacing irrelevant decisions. v2.1.2 is a trust-recovery release based on four independent field-test reports: every fix restores confidence in something codevira already does.
| Area | What changed |
|---|---|
| Smart similarity threshold | search_decisions self-calibrat |