Structural code quality metrics for agent-written programs.

Topos gives you structural code quality metrics your agents can act on. Passing unit tests proves your code works, but Topos proves it's built to last. It measures program structure — not just syntax — giving agents concrete metrics to optimize toward on every pass. You set the target; agents handle the iteration.

The Quality Pillars

Topos evaluates code along three independent pillars:

• SIMPLE — Avoids unnecessary complexity (AST entropy & CFG cyclomatic complexity).
• COMPOSABLE — Cleanly decoupled from other modules (MDG Martin instability via GitNexus).
• SECURE — Free of dangerous API reachability and taint paths (CPG analysis).

The Medal Podium

Topos checks all three pillars and awards a Code Quality Medal based on how many pass:

Medal	Criteria
🥇 GOLD	Passes all 3 (SIMPLE + COMPOSABLE + SECURE)
🥈 SILVER	Passes 2 of 3
🥉 BRONZE	Passes 1 of 3
❌ SLOP	Passes 0 (or fails to parse)

Set your Preferences (e.g., simple,composable,secure) to tell your coding agent which pillars to prioritize when aiming for Gold under token and time budgets.

---

Quick Start

Install

# Install CLI
curl -fsSL https://docs.krv.ai/topos/install.sh | sh

Using Topos

# Preferences are how you prioritize the evaluation pillars
topos evaluate src/ -r --preferences simple,composable,secure  # classify each file in a directory

---

MCP Server

Give any MCP-compatible agent — Claude Code, Cursor, Gemini CLI, Windsurf — a live feed of Topos verdicts so it can evaluate and iterate on its own output.

Set up topos mcp in your agent

 

Step 1 — Build the dependency graph (optional but recommended)

⚠️ Recommended: Without a dependency graph, Topos cannot score COMPOSABLE — any verdict containing it (including IDEAL) is unreachable. SIMPLE and SECURE always run.

npm install -g gitnexus        # one-time per machine
cd /path/to/your/repo
topos depgraph generate        # one-time per repo; writes .gitnexus/

Re-run when imports change (new modules, renames, restructures). The cache keys on .gitnexus/ mtime and invalidates itself.

💡 Tip: Verify the binary before wiring it into editors:

topos mcp   # prints the FastMCP banner and waits on stdin. Ctrl-C to exit.

Step 2 — Register with your agent

Run from your project root — Topos auto-detects its file-access root by walking up for .git or pyproject.toml.

Claude Code

claude mcp add topos topos mcp

Gemini CLI

gemini mcp add topos topos mcp

Cursor

➕ Install topos in Cursor

Or edit .cursor/mcp.json:

{ "mcpServers": { "topos": { "command": "topos mcp" } } }

Windsurf and everything else

{ "mcpServers": { "topos": { "command": "topos mcp" } } }

Step 3 — Launch from the project root

⚠️ IMPORTANT Topos refuses to read files outside a trusted root. If you must launch from elsewhere, set it explicitly:

{
  "command": "topos mcp",
  "env": { "TOPOS_MCP_FILE_ROOT": "/absolute/path/to/repo" }
}

💡 TIP On the agent's first turn, point it at the workflow doc:

"Fetch topos://docs/workflows and follow the Topos refactor loop."

Or invoke the prompt directly: topos_refactor_until_ideal(filepath="path/to/file.py").

Smoke test

"Use topos to find the worst-scoring file in src/, propose a refactor, and verify with topos_assess_improvement."

A healthy response shows {simple: 72%, composable: 65%, secure: 95%} when GitNexus is configured. If the response is missing composable, go back to Step 1.

---

How it works

Topos measures code along the three independent quality generators and maps them to an 8-element evaluation lattice:

• SIMPLE — Built from the abstract syntax tree (AST) and control-flow graph (CFG). We calculate cyclomatic complexity of the CFG and entropy of the AST to assess complexity.
• COMPOSABLE — Built from the module dependency graph (MDG) using GitNexus, to capture inter-module dependencies. This is slightly different than the usual program dependence graph (PDG) which is used to capture intra-function dependencies. We calculate Martin Instability and Fanning metrics for the MDG to assess coupling.
• SECURE — Built from the code property graph (CPG). We calculate dangerous-API reachability and taint paths from the CPG to assess security.

graph BT
    SLOP["❌ SLOP<br/>No Medal"]
    SIMPLE["🥉 BRONZE<br/>Simple"]
    COMPOSABLE["🥉 BRONZE<br/>Composable"]
    SECURE["🥉 BRONZE<br/>Secure"]
    SC["🥈 SILVER<br/>S ∧ C"]
    SSc["🥈 SILVER<br/>S ∧ Sc"]
    CSc["🥈 SILVER<br/>C ∧ Sc"]
    IDEAL["🥇 GOLD<br/>Quality Code"]

    SLOP --> SIMPLE
    SLOP --> COMPOSABLE
    SLOP --> SECURE
    SIMPLE --> SC
    SIMPLE --> SSc
    COMPOSABLE --> SC
    COMPOSABLE --> CSc
    SECURE --> SSc
    SECURE --> CSc
    SC --> IDEAL
    SSc --> IDEAL
    CSc --> IDEAL

    style SLOP       fill:#f8d7da,stroke:#842029,color:#000
    style SIMPLE     fill:#cd7f32,stroke:#5c3a1e,color:#fff
    style COMPOSABLE fill:#cd7f32,stroke:#5c3a1e,color:#fff
    style SECURE     fill:#cd7f32,stroke:#5c3a1e,color:#fff
    style SC         fill:#c0c0c0,stroke:#4a4a4a,color:#000
    style SSc        fill:#c0c0c0,stroke:#4a4a4a,color:#000
    style CSc        fill:#c0c0c0,stroke:#4a4a4a,color:#000
    style IDEAL      fill:#ffd700,stroke:#856404,color:#000

Loading

Tip

Three Independent Pillars: SIMPLE, COMPOSABLE, and SECURE are pairwise incomparable. A file can achieve any subset of {S, C, Sc} independently. 🥇 GOLD is the intersection of all three.

Manager Priorities & Agent Iteration

In a perfect world, every file would earn a 🥇 GOLD medal. In reality, managers and developers have a finite budget of time and tokens.

Topos allows you to set Preferences — an ordering of these medals based on your immediate priorities. Coding agents use this ranking to aim for 🥇 GOLD. If achieving 🥇 GOLD isn't feasible within the budget, the preference ranking tells the agent exactly how to relax its goals, ensuring it still delivers the highest possible quality medal aligned with your priorities.

---

Contributing

Topos is used internally at Krv Labs to manage AI agent code output. We welcome bugs, ideas, and contributions.

• Bug? Open an Issue
• Idea? Start a Discussion or open a PR
• Collaborate? team@krv.ai

---

Full Documentation · Measures & Metrics · Category Theory Concepts

Built with ❤️ by Krv Labs