Conversation
Co-authored-by: pelikhan <4175913+pelikhan@users.noreply.github.com>
- Add --mcp-debug flag to claude_args for improved MCP debugging capabilities - Render allowed tools list as YAML comments above the with: section for better visibility - Fix claude_args formatting to properly handle standalone flags vs paired arguments - Optimize comment line appending using gosimple suggestion Both changes improve the debugging and visibility of Claude Code Action v1.x workflows while maintaining backward compatibility and all test coverage. Co-authored-by: pelikhan <4175913+pelikhan@users.noreply.github.com>
Both changes have been implemented successfully:
The implementation properly handles standalone flags vs paired arguments in the claude_args formatting, and all tests pass. The changes improve debugging capabilities while maintaining full backward compatibility. Commit: faa0f9c Warning Firewall rules blocked me from connecting to one or more addresses (expand for details)I tried to connect to the following addresses, but was blocked by firewall rules:
If you need me to access, download, or install something from one of these locations, you can either:
|
|
@copilot add playwright tool in dev agentic workflow |
Co-authored-by: pelikhan <4175913+pelikhan@users.noreply.github.com>
Added playwright tool to the dev agentic workflow with GitHub domain access. The workflow now includes containerized Playwright support for browser automation, web testing, and accessibility analysis. Commit: d16ba4a Warning Firewall rules blocked me from connecting to one or more addresses (expand for details)I tried to connect to the following addresses, but was blocked by firewall rules:
If you need me to access, download, or install something from one of these locations, you can either:
|
|
@copilot update generated code to use claude code cli directly using npx. Do not use action anymore. |
…ub Action Co-authored-by: pelikhan <4175913+pelikhan@users.noreply.github.com>
✅ Migration completed successfully! The Claude engine now uses Key changes:
All 331 tests pass and workflow compilation works correctly. You can see the new CLI structure in the recompiled Commit: 776dd9e Warning Firewall rules blocked me from connecting to one or more addresses (expand for details)I tried to connect to the following addresses, but was blocked by firewall rules:
If you need me to access, download, or install something from one of these locations, you can either:
|
|
Co-authored-by: pelikhan <4175913+pelikhan@users.noreply.github.com>
Co-authored-by: pelikhan <4175913+pelikhan@users.noreply.github.com>
Added both CLI arguments to Claude engine configuration:
Example configuration: engine:
id: claude
verbose: true
max-runs: 3Updated schemas, configuration parsing, and CLI generation. All tests pass and backward compatibility is maintained. Commit: 9ac5eb4 Warning Firewall rules blocked me from connecting to one or more addresses (expand for details)I tried to connect to the following addresses, but was blocked by firewall rules:
If you need me to access, download, or install something from one of these locations, you can either:
|
|
@copilot revert previous commit.
|
…guments - Reverted conditional --verbose and --max-runs support from previous commit - Claude now always includes --verbose flag for enhanced debugging output - --max-turns support remains available as requested - All tests pass and workflows compile correctly Co-authored-by: pelikhan <4175913+pelikhan@users.noreply.github.com>
Reverted the previous commit and implemented the requested changes:
All tests pass and workflows compile correctly with the new configuration. Commit: 8bac88d Warning Firewall rules blocked me from connecting to one or more addresses (expand for details)I tried to connect to the following addresses, but was blocked by firewall rules:
If you need me to access, download, or install something from one of these locations, you can either:
|
|
@dsyme looks like it's working. might be some log parsing issue with the verbose flag but this will make debugging mcp servers might easier to track. |
|
@copilot update claude code args to disable telemetry. set env var DISABLE_TELEMETRY=1, DISABLE_ERROR_REPORTING=1, DISABLE_BUG_COMMAND=1 Docs: Data usage
Data policiesData training policyConsumer users (Free, Pro, and Max plans): We will train new models using data from Free, Pro, and Max accounts when this setting is on (including when you use Claude Code from these accounts).
Commercial users: (Team and Enterprise plans, API, 3rd-party platforms, and Claude Gov) maintain existing policies: Anthropic does not train generative models using code or prompts sent to Claude Code under commercial terms, unless the customer has chosen to provide their data to us for model improvement (e.g. Developer Partner Program). Development Partner ProgramIf you explicitly opt in to methods to provide us with materials to train on, such as via the Development Partner Program, we may use those materials provided to train our models. An organization admin can expressly opt-in to the Development Partner Program for their organization. Note that this program is available only for Anthropic first-party API, and not for Bedrock or Vertex users. Feedback using the
|
| Service | Anthropic API | Vertex API | Bedrock API |
|---|---|---|---|
| Statsig (Metrics) | Default on.DISABLE_TELEMETRY=1 to disable. |
Default off.CLAUDE_CODE_USE_VERTEX must be 1. |
Default off.CLAUDE_CODE_USE_BEDROCK must be 1. |
| Sentry (Errors) | Default on.DISABLE_ERROR_REPORTING=1 to disable. |
Default off.CLAUDE_CODE_USE_VERTEX must be 1. |
Default off.CLAUDE_CODE_USE_BEDROCK must be 1. |
Anthropic API (/bug reports) |
Default on.DISABLE_BUG_COMMAND=1 to disable. |
Default off.CLAUDE_CODE_USE_VERTEX must be 1. |
Default off.CLAUDE_CODE_USE_BEDROCK must be 1. |
All environment variables can be checked into settings.json (read more).
Co-authored-by: pelikhan <4175913+pelikhan@users.noreply.github.com>
Added telemetry disabling environment variables to Claude CLI execution. The following variables are now set for all Claude workflows:
This ensures no data is sent to external telemetry services for privacy and security. Commit: 909eb9d Warning Firewall rules blocked me from connecting to one or more addresses (expand for details)I tried to connect to the following addresses, but was blocked by firewall rules:
If you need me to access, download, or install something from one of these locations, you can either:
|
pelikhan
changed the title
…P tools - Add GetCopilotAgentPlaywrightTools() function to shared agentic engine code - Update Claude engine to expand playwright tools to include all 21 copilot agent tools - Update Codex engine with same playwright tool expansion functionality - Add comprehensive test coverage for playwright tool expansion - Update neutral tools tests to verify expected playwright MCP configuration - Ensure cross-engine compatibility for consistent playwright tools across all engines Implements functionality from PR #823 to generate playwright allowlist based on copilot agent supported tools, enabling full browser automation capabilities in workflows using `playwright: null` configuration. Co-authored-by: pelikhan <4175913+pelikhan@users.noreply.github.com>
Added support for generating playwright tool allowlist based on copilot agent supported tools. The implementation includes:
Now when workflows specify Commit: 3f47e10 Warning Firewall rules blocked me from connecting to one or more addresses (expand for details)I tried to connect to the following addresses, but was blocked by firewall rules:
If you need me to access, download, or install something from one of these locations, you can either:
|
| run: | | ||
| set -o pipefail | ||
| # Execute Claude Code CLI with prompt from file | ||
| npx @anthropic-ai/claude-code@latest --print --mcp-config /tmp/mcp-config/mcp-servers.json --allowed-tools "ExitPlanMode,Glob,Grep,LS,NotebookRead,Read,Task,TodoWrite,mcp__github__download_workflow_run_artifact,mcp__github__get_code_scanning_alert,mcp__github__get_commit,mcp__github__get_dependabot_alert,mcp__github__get_discussion,mcp__github__get_discussion_comments,mcp__github__get_file_contents,mcp__github__get_issue,mcp__github__get_issue_comments,mcp__github__get_job_logs,mcp__github__get_me,mcp__github__get_notification_details,mcp__github__get_pull_request,mcp__github__get_pull_request_comments,mcp__github__get_pull_request_diff,mcp__github__get_pull_request_files,mcp__github__get_pull_request_reviews,mcp__github__get_pull_request_status,mcp__github__get_secret_scanning_alert,mcp__github__get_tag,mcp__github__get_workflow_run,mcp__github__get_workflow_run_logs,mcp__github__get_workflow_run_usage,mcp__github__list_branches,mcp__github__list_code_scanning_alerts,mcp__github__list_commits,mcp__github__list_dependabot_alerts,mcp__github__list_discussion_categories,mcp__github__list_discussions,mcp__github__list_issues,mcp__github__list_notifications,mcp__github__list_pull_requests,mcp__github__list_secret_scanning_alerts,mcp__github__list_tags,mcp__github__list_workflow_jobs,mcp__github__list_workflow_run_artifacts,mcp__github__list_workflow_runs,mcp__github__list_workflows,mcp__github__search_code,mcp__github__search_issues,mcp__github__search_orgs,mcp__github__search_pull_requests,mcp__github__search_repositories,mcp__github__search_users" --debug --verbose --permission-mode bypassPermissions --output-format json --settings /tmp/.claude/settings.json "$(cat /tmp/aw-prompts/prompt.txt)" 2>&1 | tee /tmp/test-ai-inference-github-models.log |
There was a problem hiding this comment.
claude-code@latest seems brave as exposes us to Claude Code breaking changes. Maybe beta or latest official release ?
pelikhan
deleted the
copilot/fix-f0ab22fd-0c95-419f-b0d3-a61baae89f29
branch
Implemented support for generating the allowlist of playwright tools based on the copilot agent supported tools, as requested in comment #3290301233.
Changes Made
GetCopilotAgentPlaywrightTools()function - Located inpkg/workflow/agentic_engine.goto provide consistent access to all 21 copilot agent playwright tools across enginesexpandNeutralToolsToClaudeTools()andcomputeAllowedClaudeToolsString()to recognize playwright as a standard MCP tool and expand it with copilot agent toolsexpandNeutralToolsToCodexTools()method and updatedRenderMCPConfig()to support the same playwright tool expansioncodex_playwright_test.goand updated existing tests to verify all 21 playwright tools are properly included when usingplaywright: nullPlaywright Tools Included (21 total)
The following tools are now automatically included when a workflow specifies
playwright: null:mcp__playwright__browser_clickmcp__playwright__browser_closemcp__playwright__browser_console_messagesmcp__playwright__browser_dragmcp__playwright__browser_evaluatemcp__playwright__browser_file_uploadmcp__playwright__browser_fill_formmcp__playwright__browser_handle_dialogmcp__playwright__browser_hovermcp__playwright__browser_installmcp__playwright__browser_navigatemcp__playwright__browser_navigate_backmcp__playwright__browser_network_requestsmcp__playwright__browser_press_keymcp__playwright__browser_resizemcp__playwright__browser_select_optionmcp__playwright__browser_snapshotmcp__playwright__browser_tabsmcp__playwright__browser_take_screenshotmcp__playwright__browser_typemcp__playwright__browser_wait_forAll tests pass and workflows compile correctly with the enhanced playwright tool configuration.
✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.