In short: Ecuhawk sends results to email, chat and Jira, reads requirements from your documents, keeps tests in Git and runs in CI. Every connection uses your own accounts and goes straight from your bench to your tool; nothing passes through a Clazro server. Messages that leave the lab go only to addresses and channels you allow, and their numbers are checked against the real results first.
Chat and email
| Tool | What it does | Status |
|---|---|---|
| Email (Gmail, Outlook, any SMTP server) | Sends the run summary with the report attached, for example after a nightly run | INCLUDED AT LAUNCH |
| Microsoft Teams | Posts in a channel when a test fails or the safe state triggers | COMING SOON |
| Slack, Google Chat | The same alerts for teams on Slack or Google Workspace | COMING SOON |
Planning, defects and documentation
| Tool | What it does | Status |
|---|---|---|
| Jira | Opens a ticket for a failed test with the failure text, a trace summary and the report | COMING SOON |
| Confluence, SharePoint, Google Drive | Reads requirement and test documents from where your team keeps them | COMING SOON |
Requirements
| Source | What it does | Status |
|---|---|---|
| Text, Markdown, CSV | The AI reads the requirements in the project and drafts tests that carry their IDs | INCLUDED AT LAUNCH |
| ReqIF (Polarion, DOORS, Jama, codebeamer) | Imports requirements with their IDs for coverage tables | COMING SOON |
| PDF and scanned documents | Reads specifications that exist only as PDF | COMING SOON |
Code, CI and data
| Tool | What it does | Status |
|---|---|---|
| Git (GitLab, GitHub, Bitbucket) | Tests, bench and project files are plain text, so they are reviewed and versioned like code | INCLUDED |
| CI servers (GitLab CI, GitHub Actions, Jenkins) | A command-line run without a window; the verdict is the exit code and the report is written to disk | INCLUDED |
| JUnit XML | Test results that CI dashboards display | COMING SOON |
| Python scripts and other programs | Control the bench, run tests and read results through the same local API the app uses | INCLUDED |
| Measurement files (MF4, BLF, ASC, CSV) | Record and export traces and signals for other analysis tools | INCLUDED |
AI agents
Ecuhawk offers its bench as tools over MCP, the common protocol for AI agents. An agent such as Claude Code or Hermes Agent can then read signals, run tests, analyse the trace and draft tests, under the same rules as a person: hard limits, approval on a real bench, drafts only, fact-checked answers and an audit log. See AI and your data. INCLUDED AT LAUNCH
Rules for everything that leaves the lab
- Your accounts, direct connections. Email goes through your mail server, tickets through your Jira, messages through your chat. There is no Clazro server in the path.
- Only allowed recipients. Results go only to the addresses and channels on your allow-list.
- Checked numbers. Before an AI-written summary is sent, every number in it is compared with the values the bench actually measured. If one does not match, the message is not sent.
- Logged. Every message sent or refused is written to the audit log.
- Secrets stay local. Passwords and tokens are kept in a local settings file, never in test files or reports.
Questions
Does Ecuhawk work with Jira?
A Jira integration that opens a ticket for a failed test, with the failure text, trace summary and report, is planned. Tell us in the beta form if you need it early.
Can Ecuhawk email test reports?
Yes, this is in development: a summary with the report attached goes to the addresses you allow, through your own mail server such as Gmail or Outlook.
Can Ecuhawk post in Microsoft Teams?
Teams and Slack alerts for failed tests and safe-state events are planned.
Does our data go through Clazro when Ecuhawk sends a message?
No. Ecuhawk connects straight to your own mail server, chat and Jira with your accounts.
Can we connect a tool that is not listed?
Usually yes, through the local API, the command line or an AI agent over MCP. Tell us which tool in the beta form.