agent-config

Use when creating a Jira ticket to plan upcoming work. Guides ticket drafting with clear structure, self-contained context, and testable acceptance criteria, then creates the ticket via the available Jira/Atlassian broker tool.

Use when reviewing a pull request, branch, commit range, working tree diff, plan, document, or other coherent unit of work across correctness, security, codebase alignment, code quality, test quality, performance, and simplicity.

Use when the user asks to create or polish a standalone HTML artifact such as an HTML report, interactive explainer, visual plan, dashboard, slide deck, diagram, prototype, or throwaway HTML tool.

Use when debugging bugs, failures, exceptions, flaky behavior, regressions, or performance problems where the cause is not already proven.

Use when the user asks to build web components, pages, or applications that need distinctive, production-grade frontend interfaces with high design quality

Reference skill for browser automation: web testing, form filling, screenshots, and data extraction. Use when the user needs to navigate websites, interact with web pages, fill forms, take screenshots, test web applications, or extract information from web pages.

Use when implementing a feature or bugfix that involves writing meaningful application logic

Use when turning an implementation plan, project plan, design plan, roadmap, migration plan, or execution checklist into a visual HTML document for easier human review.

Use when designing, building, debugging, or reviewing AI coding agent harnesses — single-agent shape (tools, prompts, context, hooks, model selection) or multi-phase workflows (orchestration, subagents, verifiers, ticket-to-PR pipelines). Covers model-specific guidance for Claude 4.x and GPT-5.x families, and platform-specific patterns for Claude Code, the Claude Agent SDK, and Pi. Invoke when the user asks about harness design, scaffold patterns, agent loops, subagent orchestration, verification strategy, context compaction, plan/implement/verify pipelines, or how a particular model changes harness choices.

Use when requirements, scope, product behavior, edge-case policy, acceptance criteria, or design intent are fuzzy and need an interview before planning or implementation.

Use when turning a clarified implementation request, feature idea, bugfix, refactor, workflow change, or design discussion into an execution-ready plan that a fresh coding agent can implement autonomously, especially before running /goal.

Use when creating, editing, validating, manually running, or debugging Pi scheduled task Markdown files managed by the scheduled-tasks extension.

Use when creating a Jira ticket to plan upcoming work. Guides ticket drafting with clear structure, self-contained context, and testable acceptance criteria, then creates the ticket via the broker-backed Atlassian MCP tool.

Use when reviewing a pull request or branch holistically across multiple dimensions (correctness, security, codebase alignment, code quality, test quality, performance). Accepts a PR URL or branch name as argument.

Use when ingesting external content (Jira, Confluence, GitHub, docs, web pages) into Hindsight memory, retaining facts for long-term recall, or querying stored memories via Hindsight's MCP tools. Covers stable document IDs, tagging conventions, retain/recall/reflect selection, and how to avoid duplicate memories.

Automates browser interactions for web testing, form filling, screenshots, and data extraction. Use when the user needs to navigate websites, interact with web pages, fill forms, take screenshots, test web applications, or extract information from web pages.

Use when implementing a feature or bugfix that involves writing meaningful application logic

Use when finishing the structured development workflow after verifying work - cleans up plan files, reflects on learnings, and presents PR options

Use when creating a new skill or updating an existing skill

Use when stress-testing, challenging, reviewing, repairing, or grilling a plan before execution.

Use when the user asks Claude Code to create a standalone HTML artifact, HTML report, interactive explainer, visual plan, dashboard, slide deck, diagram, or throwaway HTML tool.

Use when you have a spec or design document and need to break it into a detailed implementation plan with right-sized tasks

Use when you have a written implementation plan file to execute - dispatches subagents for implementation and reviews to prevent context pollution

Use when the user asks to build web components, pages, or applications that need distinctive, production-grade frontend interfaces with high design quality

Use when creating a new skill or updating an existing skill

Use when designing a significant feature that needs requirements exploration and upfront design before implementation

Use when troubleshooting a system issue, investigating an outage, debugging a production problem, or responding to an incident

Use when finishing plan execution to holistically review all work before completing - runs automated checks, dispatches 5 parallel reviewers, auto-fixes clear issues, surfaces ambiguous findings

Use when running open-ended, multi-session technical research that requires experiments, evidence gathering, and an accumulating HTML report

Use when you have a written implementation plan file to execute quickly - does implementation and reviews inline in main context without subagent dispatch