Agent Updates

Author: strawgate

.github/workflows/claude-on-mention.yml

@@ -43,6 +43,8 @@ jobs:
           2. Get the ${{ github.event_name == 'issues' && 'issue' || 'pull request' }} ${{ github.event.issue.number || github.event.pull_request.number }} in the GitHub repository: ${{ github.repository }}.
           3. Don't forget about your MCP tools to call search_code, get_files, etc. to search the repository and other repositories to identify the related classes, methods, docs, tests, etc that are relevant to the code.
           4. Be Thorough! Go the extra mile! Do great work!
+          5. If anything you needed or wanted to do during your work was not possible, document the problem in a `Problems Encountered` section
+          of your response. Especially if it was something you were asked to do like research the web, run tests, lint, etc.
 
           # Using Make and Git Commands
           You can run make commands (e.g., `make lint`, `make typecheck`, `make sync`) to build, test, or lint the code. You can also run git commands (e.g., `git status`, `git log`, `git diff`) to inspect the repository. You cannot run arbitrary bash commands - only make and git commands are allowed.

.github/workflows/claude-on-open-label.yml

@@ -57,6 +57,7 @@ jobs:
             3. Provide an high quality and detailed plan that a junior developer could follow to implement the recommendation
 
           Populate the following sections in your response:
+            Problems Encountered during the investigation
             Recommendation (or “No recommendation” with reason)
             Findings
             Detailed Action Plan
@@ -68,6 +69,12 @@ jobs:
 
           When formulating your response, you will never "bury the lede", you will always provide a clear and concise tl;dr as the first thing in your response. As your response grows in length you can organize the more detailed parts of your response collapsible sections using <details> and <summary> tags. You shouldn't put everything in collapsible sections, especially if the response is short. Use your discretion to determine when to use collapsible sections to avoid overwhelming the reader with too much detail -- think of them like an appendix that can be expanded if the reader is interested.
 
+          # Example of Problems Encountered during the investigation:
+          
+          While performing the investigation I encountered the following problems:
+            - A tool that I thought was available I was not able to use because ...
+            - I was not able to run tests because ...
+
           # Example output for "Recommendation" part of the response
           PR #654 already implements the requested feature but is incomplete. The Pull Request is not in a mergeable state yet, the remaining work should be completed: 1) update the Calculator.divide method to utilize the new DivisionByZeroError or the safe_divide function, and 2) update the tests to ensure that the Calculator.divide method raises the new DivisionByZeroError when the divisor is 0.
 

AGENTS.md

@@ -234,3 +234,8 @@ make bump-version-dry VERSION=1.2.3    # Dry run
 - For human developer documentation, see [DEVELOPING.md](DEVELOPING.md)
 - For library usage documentation, see [README.md](README.md)
 - For package-specific information, see READMEs in each package directory
+
+## Radical Honesty
+
+Agents should be honest! Properly document any problems encountered, share
+feedback, and be transparent about your AI-assisted work.