DVERA (Dynamic Verification Agent) is a set of coding-agent skills for CT (Controller Tester) C/C++ verification workflows, maintained by Suresoft Technologies. It brings CT verification work into coding agents such as Claude Code, Codex CLI, Cursor, and GitHub Copilot, using work context such as source code, requirements, build information, source changes, and existing test assets to prepare and initiate CT verification work.
CT (Controller Tester) is a test automation solution for mission-critical C/C++ software, covering unit testing, integration testing, and code-based testing, with integrated test environment setup, test design and generation, execution, code coverage analysis, reporting, and traceability management. DVERA does not replace CT's verification results: coding agents help prepare and coordinate verification work, while CT handles analysis, builds, test execution, and result aggregation. Generated tests must be reviewed for test intent, data, and expected results, and then confirmed with actual execution results and measured coverage.
In terms of usage, the skills in this repository document verification workflows that connect to the product and are not themselves a standalone replacement for a CT installation. After installation in Claude Code or via the terminal, the skills are loaded as commands such as /dvera:ct-init-project and /dvera:ct-test-loop. The plugin adds approximately 620 tokens per session, and each skill body is read only when that skill runs. Productized operation requires a deployed CT environment; without it, the skills stop at the installation check and provide guidance on how to contact product support.
The workflow overview includes: preparing compilers, macros, include paths, toolchains, and target context for a CT project; analyzing the project and resolving environment or configuration issues through a controlled feedback loop; generating, executing, and improving unit or integration tests using build, execution, and coverage feedback; reviewing coverage and results, generating reports, and rerunning affected verification after source changes. The repository provides trigger examples, but responses may vary depending on the AI model and conversation context; if results do not match expectations, you can add details and retry.
Productized workflows depend on a deployed CT and DVERA environment. This repository documents the verification workflow and includes an MCP server that forwards tool calls to a local CT installation, but it does not include CT runtimeor product installation. When there is no deployed environment, users can view the workflows recorded in the repository, but skills requiring CT operations will stop at the installation check and must not claim that CT analysis, execution, coverage measurement, or reporting has occurred.
The MCP server (scripts/dvera-mcp.py) lets agents use CT's verification tools directly. Once connected, agents can create projects, run analysis, generate and execute tests, and read coverage without leaving the conversation; these are the same tools CT exposes to its own agents. Calls go to CT's ct_tool.py entry point, and the server does not retain its own process between calls. There are two ways to connect: first, point the MCP client at CT's bundled interpreter, with no additional installation required; on Linux the interpreter is located at <CT install>/python/python3, and any Python 3.8 or higher can also be used, with the script path replaced by the location of this repository (either a self-cloned copy or a copy installed via a plugin manager; claude plugin list can show the plugin location). The server obtains the source directory from the client's working directory, so the client must be run in the project directory to be verified. Second, download dvera-mcp.mcpb from this repository's releases and open it with a client that supports installing MCP bundles; it resolves Python on its own, with no prior installation required. If CT is installed outside the default location, set CT_HOME. The server can safely retain its configuration on any machine: it can start whether or not CT is installed; without CT it provides only one tool, dvera_check_environment, which reports missing items and how to address them; with CT that tool is also present, and it should be run first when the verification tools do not appear.
Requirements for productized verification include: CT 2026.06 or later; a valid CT license; a supported C/C++ build and target environment; and the DVERA integration provided as part of a deployed CT environment. For product demos, purchases, or deployment inquiries, contact bizcenter@suresofttech.com.
The MCP server in this repository runs entirely on the user's local machine and sends nothing to Suresoft Technologies or any third party; there is no telemetry, analytics, or network calls of its own. Tool parameters are passed to the CT installation on the same machine, and CT's responses are returned to the MCP client, without writing anything except CT's own workspace. It does not store credentials, and it does not read files outside the CT installation directory and the paths specified by tool calls. Anything CT itself records (workspace, analysis output, reports) is governed by CT's own data handling, not this repository. Skills are documentation and do not run any code on the user's machine. See PRIVACY.md for the full policy; questions about CT's own product data handling can be sent to bizcenter@suresofttech.com.
Regarding licensing and trademarks, this repository uses the MIT License, covering all content in the repository, including documentation, skill files, and MCP server source code, but not CT itself: CT and DVERA are products of Suresoft Technologies, and this repository grants no product functionality or product license. GoogleTest is a trademark of Google LLC; this repository is not affiliated with, sponsored by, or endorsed by Google; other product names may be trademarks of their respective owners.
