Today has all been about learning new skills. But in this moment I wanted to talk specifically about AI skills.   During the course of today I attended an excellent AI chapter event on AI skills and I also decided to do a little bit of experimentation for myself.....

At work over the last few weeks I have been using Claude with some new skills that one of the developers on my team has created.   These skills allow us to create and merge documentation from various sources and then publish into final PDFs for inclusion in release documentation.

For a little while I had been intending to try out the new skill Simon Tomes created as part of an insights article on an Exploratory Testing Charter Planner (ETCP). This provided a Claude skill that could be used to assist with the definition of an exploratory testing charter. 
https://www.ministryoftesting.com/insights/the-exploratory-testing-charter-planner-etcp

Finally today I found some time to have a look at in more detail and start looking at what gets included in a skill.  However it just so happened today that there was also a Ministry of Testing AI charter event with Maneesh Maddala and Rahul Parwal that covered this very topic!   This was a really excellent session with lots of useful information and ideas on applying skills in AI to our daily work as developers and quality engineers.   

After work today I decided to get the exploratory testing charter planning skill onto my system and take it for a spin.....

It was nice and easy to set the skill up on my machine.   (Actually I will let you into a little story...after downloading it I was having a bit of trouble getting it installed correctly as the file had a .skill extension which didn't see to be recognizable to Claude...turns out I had to rename it to a .zip file and problem solved!!).  I just needed to unzip the skill into my local .claude directory and my skill was in place!!.   I spun up a Claude session in the terminal and there was the skill all ready to go!!

So first up when I ran the skill it set the scene on what it could help me with and the steps involved e.g. an exploratory testing charter. 

Then it gradually stepped me through the various components of the charter.   We started off working out the charter name.  As an experiment I decided to ask Claude to access JIRA via the MCP server and suggest a charter name.  This worked out quite nicely, I reviewed and then refined my test charter.

Next up we went through working out the various other components of the charter, so we worked through framing Oracles and Risks (in each case I asked Claude to give me ideas for those and of course refer for useful information in the JIRA ticket).  

Finally we worked through questions I wanted answered during the session followed by determining heuristics (In this case I suggest the JIRA ticket itself, my expectations of what the JIRA ticket was intended to fix).  This helped flesh out some really useful heuristics for testing this fixed issue.  

We wrapped up the process by considering any environment and setup I needed and defined a test session timebox.   There was also a final question about any tips I wanted to add to the exploratory testing charter.

The complete process ended with the creation of a markdown document that documented my test charter.  Along the way there are some useful links on external and Ministry of Testing resources to guide you in working all the elements of the charter.

For me the next steps will be interesting.  I intend to try out a bit of an experiment using Claude as my testing buddy for the session I created...so for example during the 90 minutes capture on the console what I've tested, note down my observations,  note down test ideas or things to look more closely at and maybe ask it for additional test ideas.   The idea with this experiment will do the testing I need to over the allocated timebox and then at the end of the session get it to help me me review and summarise my findings.

I did briefly ask Claude to look at the testing charter I had produced and give me some starter ideas for my testing and it provided me some very useful initial ideas!! So that is an encouraging start...

I personally think with AI there is a use for this tooling in our jobs as quality engineers for it to assist us in the process of exploratory testing.  I'm certainly not suggesting this be a replacement for critical thinking and doing good quality testing but a tool to assist in that task and help me summarise my findings at the end of the session.  One of the things I find quite hard is summarising my session at the end so I'm hoping that AI can help me collate my findings, ideas and outcomes. Neither do I think this is a replacement for talking though the outcomes of a testing session with another quality engineer or developer on your team.     

It was really cool today to attend a Ministry of Testing AI chapter event and then then do a bit of exploring of skills myself.