Enhance report generation, UI settings, and request analysis capabilities - #23
Merged
Conversation
- Added HTML report generation in createRunReporter and writeReport functions. - Implemented writeRunHtml to create HTML reports for benchmark runs. - Updated tests to verify HTML report creation and viewer functionality. - Introduced sanitizeReportForViewer to limit bulky stdout in reports. - Added new scripts for viewing benchmark results in package.json.
- Updated profiles.ts to include UiSettingsSnapshot in profile options and methods. - Enhanced protocol.ts to add ui property in SettingsProfileView interface. - Implemented reviewWorkspaceFile message handling in sidebar.ts for workspace file reviews. - Introduced ModelQuickSelect component in App.tsx for model selection with custom input. - Modified ReviewPanel.tsx to improve file review UI and interaction. - Adjusted CSS styles for improved layout and responsiveness in various components. - Updated RELEASE.md to clarify npm authentication requirements and token usage. - Enhanced publish-npm script to provide guidance on handling EOTP errors in CI.
…-ups with high confidence
…eact.jsonl, update suite.json for new structure - Deleted medium.jsonl and its associated test cases. - Introduced monorepo.jsonl with new test cases for shared and API packages. - Added react.jsonl with unit tests for React components and hooks. - Updated suite.json to reflect the new case files and adjusted expected counts. - Modified load.test.mjs to validate the new case structure and counts.
… bands - Introduced `index.html` for the Policy Admin UI with budget and re-read settings. - Created `preview.js` for client-side budget calculations and formatting utilities. - Implemented `server.mjs` to serve the HTML interface and handle API requests for saving policy configurations. - Added `shipWriter.mjs` to write policy band configurations to TypeScript files. - Established a structure for managing compact, standard, and wide policy bands with appropriate overrides.
- Added a new function to categorize settings for verify, files, and skills. - Updated to display setting vs live counts in a split box format. - Introduced a function to show derived metrics in the live preview. - Improved the layout and descriptions in for better clarity and usability. - Adjusted labels and hints throughout the UI for consistency and enhanced user experience.
- Introduced a hierarchical context usage node structure for better token budgeting. - Updated the prompt_ready event to include detailed budget information for token meters. - Enhanced the TokenMeter component to visualize the token budget tree, including allocated and used tokens. - Implemented merging of prompt budget data into the context usage breakdown. - Added tests for context usage breakdown and merging functionality. - Improved styling for the token context slices in the UI.
…line coverage and continuation support - Updated read_file output to include startLine, endLine, totalLines, eof, nextStartLine, and truncationReason. - Modified executeReadFile and executeReadManyFiles actions to handle maxContentChars for character budgeting. - Introduced line-window selection logic in readFileWindow for efficient content retrieval based on line and character limits. - Enhanced InMemoryFileSystemAdapter and NodeFileSystemAdapter to support new readFile options and return detailed results. - Added tests for read_file functionality to verify coverage cursors and continuation behavior. - Updated schemas and contracts to reflect new output structure and options for file reading.
… + GitHub, Delivery bus, Log / incident product. VS Code + polis, - Sentry (+1 APM), CI log puller, evidence packs, ticket templates with full logs, patch attach, idempotent issues. - Event ingress, webhook server, GitHub event adapters, workflow_run failure → triage agent. Native create_issue / create_pull_request tools (or MCP). Artifact packing. Dedupe fingerprints. - Unify connect adapters as delivery targets. Slack/Discord/Telegram schedule results. GitHub Check Runs + PR comments. Outbound signed webhooks. Delivery retries + receipts. - Wire origin through SDK → intake → Decision Policy. Autonomy presets. Exit codes. --agent / --prompt-file. Official GHA composite action + 2 workflow templates (post-commit, CI-fail). Skills: cicd-agent, incident-triage. Docs cookbooks.
- Introduced a new event specification for post-commit cover automation. - Created smoke test scripts for post-commit cover and CI failure scenarios. - Enhanced incident evidence handling by redacting sensitive information. - Implemented GitHub webhook signature verification for enhanced security. - Added functionality to prevent pushes to protected branches (main/master). - Updated tooling to support idempotent issue creation and commenting in GitHub. - Improved test coverage for security features and Git push validation.
- Add `buildStallContinueRationale` and `shouldOfferStallContinue` functions to summarize exploration stalls and determine if continuation is needed. - Update action exports to include the new stall rationale functions. - Create unit tests for stall rationale and decision functions. - Enhance constants to support new decision types related to grant expansions and continuation requirements. - Modify input and output schemas to accommodate new decision structures. - Update the execution pipeline to handle grant expansion and continuation scenarios, including rationale generation and checkpoint management. - Refactor decision policy to recognize large implementation tasks and adjust planning depth accordingly. - Adjust prompt construction to account for plan budget tokens and integrate plan sections into the prompt pipeline.
…ecision policy - Introduced additional golden decision cases in to enhance testing coverage. - Created a new file to generate pairwise-style matrix cases for various intents, modes, and scopes. - Updated to include guidance for using web search and fetching URLs based on granted tools. - Enhanced to test the behavior of the model when instructed to call web search. - Exported function from for broader accessibility. - Updated to link the package for workspace management.
…ding providers and related interfaces
- Added support for explicitly attached skills via @skill: mentions and CLI options. - Introduced in various schemas to manage required skills. - Implemented functions to parse skill mentions and merge them with explicit IDs. - Enhanced SkillsPipeline to resolve required skills and report omissions. - Updated tests to cover new functionality for required skills and their resolution.
- Added detailed guidance for mutation tools in BuildSystemAndConversation.ts to clarify write authority and retry logic. - Modified node-express index.js to conditionally bind the server port only when the file is the entry point, preventing hanging during tests. - Updated mitii-benchmark-agent.mjs to handle SIGTERM/SIGINT signals gracefully, ensuring structured output is emitted even on timeout. - Introduced rerun-failed-subset.mjs script to facilitate re-running a fixed set of failed test cases with workspace retention for inspection. - Enhanced runner.mjs to save full agent transcripts for post-mortem analysis when workspaces are kept. - Updated express.jsonl test cases to ensure proper formatting and coverage for various endpoints and scenarios.
…ll id allocation in GeminiLlmPort
- Added tests for effective context window resolution and fallback behavior. - Implemented normalization for maximum output tokens, mapping legacy values to auto. - Updated sidebar settings persistence to skip unchanged provider settings. - Introduced model refresh handling in the app state with appropriate UI feedback. - Enhanced settings panel to support model refresh and display normalized maximum output tokens. - Updated protocol to include settings saved acknowledgment. - Added timeout handling for provider model listing to improve connection reliability. - Refactored window budget policies to scale compaction ceilings with context window size. - Adjusted default window budget settings for improved performance across various contexts.
…ath normalization
- Added new types and options for RequestUnderstandingPipeline, including RequestUnderstandingOptions. - Implemented resolveIntentClassifierMaximumOutputTokens to dynamically set maximum output tokens based on context window size. - Introduced isWholeRequestReadOnlyConstraint to detect read-only requests, improving intent classification. - Enhanced RuleIntentClassifier to utilize isWholeRequestReadOnlyConstraint for better handling of constraints. - Added resolveFuzzyFileTargets to improve file target resolution based on candidate paths. - Updated TaskAnalyzer to utilize candidateRelativePaths for fuzzy resolution of file targets. - Enhanced tests for new functionalities, including isWholeRequestReadOnlyConstraint and resolveFuzzyFileTargets. - Improved SelectProportionalChecks to omit browser/e2e tests when not required, ensuring more relevant checks are selected. - Updated verification policy to prioritize unit tests over browser/e2e tests in script candidates.
codewithshinde
added
enhancement
…handling in automation message processing
…and improve test coverage for markdown creation grants
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This pull request introduces significant improvements to CI/CD automation and documentation for Mitii, focusing on enabling unattended agents for CI workflows, adding channel bridge documentation, and updating Node/npm requirements for publishing. The main changes include new GitHub Actions for running Mitii agents in CI, detailed documentation for the new
connectfeature (Telegram, Discord, Slack), and updates to workflows for trusted npm publishing and compatibility with Node.js 22.CI/CD Automation Enhancements:
.github/actions/mitii-runto run Mitii headless agents in CI, supporting flexible prompts, agent selection, and autonomy levels..github/workflows/mitii-post-commit-cover.yml: Runs a post-commit coverage agent after pushes to main branches, auto-generates tests, and opens draft PRs if tests pass..github/workflows/mitii-ci-failure-triage.yml: On CI failure, triages logs, creates or updates a GitHub issue, and may attach a patch using a Mitii agent.mitii-agent.ymlworkflow to point users to the new automation workflows. [1] [2]Documentation and Developer Experience:
apps/cli/README.mdwith comprehensive instructions for the newmitii connectfeature, detailing setup and usage for Telegram, Discord, and Slack channel bridges, including troubleshooting and security best practices.Build and Publishing Pipeline Updates:
.github/workflows/npm-publish.ymlto use Node.js 22, install npm 11.5.1, and support GitHub OIDC trusted publishing for npm. [1] [2]release.ymlfor compatibility with the updated npm publish workflow. [1] [2]Miscellaneous Improvements:
README.mdto 2.9.7..vscode/settings.json.These changes collectively modernize Mitii's CI/CD automation, add robust unattended agent support, and provide clear documentation for new features and requirements.