
Plan and coordinate parallel subagent lanes using OpenCaw roles, durable task artifacts, and tool-specific execution guidance.
Run or structure database connectivity and query workflows through supported CLI tools for MSSQL, MySQL, PostgreSQL, SQLite, Azure Storage Tables, Cosmos DB metadata, and Databricks SQL.
Create a detailed task folder and TASK.md, then create/link a matching GitHub issue and track its URL while open.
Install or preview installation commands for database CLI tooling across MSSQL, MySQL, Cosmos DB, Azure Storage Tables, SQLite, PostgreSQL, and Databricks.
Create and maintain one GitHub issue per task, track only open issue URLs in `.ai/tasks`, and remove closed links during sync.
Record a user correction as both a memory pattern and a preventive rule in the host repository.
Promote a durable project fact into the correct host repository fragment file outside the mounted baseline.
Build the host .NET solution or project and summarize the first meaningful compiler error if it fails.
Ensure task checklist files exist, keep issue URLs for open work only, and sync out closed issue links.
Generate non-interactive Playwright evidence reports that replace Playwright UI mode and HTML report browsing in OpenCaw sessions.
Ensure pull requests are explicitly associated with the task issue using GitHub closing keywords.
Post QA and Playwright verification evidence to the linked task issue, including result summary and screenshot references.
Discover live web UI structure with Playwright or playwright-cli before authoring browser tests or page objects.
Design, implement, and validate Playwright end-to-end coverage for critical user workflows.
Refine, rerun, and stabilize Playwright tests using prior failures, traces, screenshots, and source-controlled test code.
Run task QA immediately after a confirmed PR is available, then post pass/fail evidence to the PR with inline screenshot references.
Stop before pushing or opening a pull request, summarize readiness, and ask the human to confirm they are ready for PR creation.
Review API Management related changes and summarize policy, routing, versioning, and deployment considerations.
Archive a host repository task note into `.ai/archive/tasks` for traceability.
Review Azure-related changes and list required environment settings, identity assumptions, and deployment considerations.
Suggest a branch name using division conventions for the requested work item.
Clean, restore, and rebuild the host .NET solution.
Review package usage in the host .NET repository and identify likely update, duplication, or cleanup opportunities.
Enforce text language and safety policy for generated images with explicit user-language override handling.
Run dotnet format against the host repository.
Generate the host repository ARCHITECTURE.md by composing one or more selected templates from ./.architecture/.
Iterate generated images until they pass mandatory sanity rules for anatomy, text quality, language alignment, and safety.
Evaluate whether a discovered fact is durable enough to promote into host repository memory.
Inspect the host .NET repository and produce a concise project map, responsibilities, and common commands.
Review host repository changes that touch Azure Service Bus and summarize sequencing, configuration, and retry considerations.
Review Azure storage-related changes and summarize account, container, blob, and SAS or identity considerations.
Execute and verify .NET runtime/framework upgrades with minimal behavioral drift.
Run tests for the host .NET repository and summarize failures clearly.
Restore NuGet packages for the host .NET solution or project.
Record a reusable bug diagnosis or resolution in the host repository debug history.
Convert a .NET bug report or feature request into an ordered implementation plan with validation steps.
Generate a concise conventional commit message from the current diff and commit host repository changes.
Review Azure Function changes and summarize trigger, configuration, binding, and deployment considerations.
Create the recommended host repository .ai scaffold and ensure host AGENTS bootstrap points to OpenCaw when missing.
Clean and compress host-repository context by compacting completed task files, deduplicating memory/rules/debug notes, and refreshing a high-signal context summary. Use after finishing tasks, before handoff, or when context artifacts become noisy.
Manage explicitly requested OpenCaw goals as automated multi-task delivery flows with automatic per-task PR creation and required post-PR QA.