etc-nixos

Deeply gather context from an Azure DevOps work item by following its relations, linked PRs, hyperlinks, and comments. Use whenever a command needs the full picture behind a work item -- planning features, reviewing PRs, triaging, or estimating.

Common steps for creating a new Azure DevOps work item in the FundGuard project, including iteration assignment, parent linking, and triage. Called by create-task, create-bug, and request-environment-access skills — not invoked directly by the user.

Mechanical estimate-and-transition of an Azure DevOps work item to the Triaged state. Called programmatically by skills that create or triage work items -- not invoked directly by the user. For user-initiated triage, use the triage skill instead.

Transitions an Azure DevOps work item to the Active state. Use when starting work on a work item, given an explicit ID or one inferred from the current feature branch.

Marks an Azure DevOps work item as Blocked and links it to its predecessor blocker. Use when a work item cannot proceed until another work item is completed.

Investigates bugs to find root cause, designs a TDD-based fix plan, and implements after approval. Use when given a bug ticket or description of defective behaviour that needs investigation and fixing in the codebase.

Follows up on a previous /review-pr or /review-plan review to check whether findings were addressed. Use in the same conversation as a preceding review invocation.

Performs a structured code review on a pull request and posts feedback. Use when given a PR ID, inferred from the current branch, or extracted from a Slack message.

Resolves the environment and organisation from an Azure DevOps work item, looks up the full environment name on Sunday, constructs the Internal Gateway URL, and sets INTERNAL_GATEWAY in the @fg/gql-api .env file. Use when given a ticket ID or URL to configure the internal gateway for local development.

Finds human-authored PR comments, replies with governing artifact citations, and suggests /create-task for uncovered gaps. Use when given a PR ID, inferred from the current branch, or in the same conversation after /review-pr.

Suggest and apply improvements to agent skills, rules, and subagent prompts after execution. Use proactively after running any skill or subagent -- especially when the execution required workarounds, discovered new information, or hit outdated instructions.

Fetches Slack thread context, researches the topic using agent skills and MCP tools, and presents findings for the user to decide on a reply. Use when given a Slack message permalink to understand and respond to a conversation.

Records an out-of-scope fix as a blocked ADO work item linked to the current task and adds a TODO comment in code. Use when deferring a fix discovered during current work so it remains traceable without expanding scope.

Produces a proposed sequence of commits to ship work in a single PR, including behaviour-preserving refactoring commits when needed. Use when planning a feature, bug fix, or task from a ticket, textual description, or tech design.

Deeply investigates an Azure DevOps work item — gathering context, examining the codebase, and proposing a fix approach — before drafting ADO field updates for user approval. Use when triaging a work item with codebase-informed investigation, unlike the mechanical triage-transition skill.

Catalogue of recommended event-driven automations with portable skill references and platform compatibility guidance. Use when setting up background automations on any platform (Cursor, Claude Code, Antigravity) or when deciding which automations to activate for a project.

Analyzes branch or PR code changes to identify opportunities for new or evolved agent artifacts (rules, skills, subagent prompts) in the target repository. Suggestions are grounded in the actual work just completed. Use after significant implementation work -- especially before submitting a feature for review.

Opens a pull request, links it to the feature work item, transitions the item to Code Review, and notifies the team on Slack. Use when submitting a feature for review, given a feature ID or a feature/<id> branch name.

MANDATORY rules for all external messages -- approval workflow, voice and tone, formatting, post-action linking. MUST be read in full before composing ANY text for Slack, PR comments, work-item updates, or other external platforms. Content evolves; never assume from memory.

Framework for investigating production incidents. Use whenever the agent analyzes alerts, outages, or performance degradations.

Communication principles derived from Peikoff's epistemological framework -- motivation, delimitation, structure, concretisation, self-containment, objectivity, and anti-rationalism. Use whenever the agent composes text that will be committed, posted, or published -- PR descriptions, commit messages, Slack messages, plans, technical designs, PRDs, documentation, READMEs, code comments, work-item descriptions, review comments, wiki pages. Do NOT use for agent-to-user conversation in the IDE.

Common steps for creating an Azure DevOps pull request, optionally linked to a work item, with user-approved title and description. Called by submit-feature and other skills that need to open a PR — not invoked directly by the user.

Verifies a merged PR and resolved work item, unblocks dependents, and cleans up the git worktree and branches. Use after a feature or hotfix PR is merged and the worktree should be torn down.

Performs codebase reconnaissance and produces a structured effort estimate in hours with confidence level. Use when given an Azure DevOps work item ID or free-text task description that needs effort estimation before planning or sprinting.

Safely manage API evolution, especially breaking changes. Plan mode for design, then Agent mode for implementation. Covers impact assessment, expand-and-contract pattern, consumer migration, endpoint deprecation, and validation. Based on Newman's emphasis that independent deployability lives or dies at the API boundary.

Gathers observability data from Datadog for a production incident and produces a structured investigation summary. Use when given an incident ID or a description of symptoms that need investigation.

Composes and posts a pipeline bypass request to the

Classifies and creates or amends agent artifacts (skills or subagent prompts) under home/file/agents/ from a behavior description. Use when adding new agent behavior to the dotfiles repository that deploys to ~/.agents/, ~/.cursor/, ~/.claude/, and ~/.gemini/.

Verifies that a skill or workspace rule stays portable across AI agents. Provides a portability checklist for SKILL.md files and applies the portable/generated/agent-specific trichotomy for workspace rules. Use whenever creating or modifying skills, rules, or subagent prompts.

Guides creation of workspace-level rules that work across multiple agents. Applies the portable/generated/agent-specific trichotomy to decide where rule content lives and how it reaches each agent. Use whenever the agent creates or modifies workspace rules in any repository.

Commit hygiene and structuring guidelines. Use whenever the agent creates git commits.

Build a new microservice from scratch, ensuring it meets all "good citizen" standards from Newman's Building Microservices. Plan mode for design, then Agent mode for implementation. Covers scope definition, boundary validation, API design, data ownership, workflows, deployment, good-citizen checklist, security, testing, and observability.

Reviews recent agent transcripts for recurring friction, workarounds, and failures, then proposes batch improvements to skills, rules, and subagent prompts. Use on-demand or as a scheduled weekly automation to keep the skill ecosystem evolving from evidence.

Diagnoses a failed Azure DevOps pipeline build and presents a structured summary of errors, failed tests, and changes. Use when given a build ID or when investigating the latest build failure on the current branch.

Architectural thinking principles from "The Software Architect Elevator" (Hohpe) -- options thinking, rate of change, systems thinking, decision quality, communication, organisational awareness, and flow. Use whenever the agent plans features, reviews code or plans, designs artifacts, estimates work, or investigates incidents.

Design a new microservice architecture from scratch. Plan mode workflow covering domain modeling, service boundaries, communication, workflows, deployment, observability, security, resiliency, scaling, and validation against Newman's principles.

Extract a microservice from a monolith or existing system. Plan mode for analysis, then Agent mode for implementation. Covers goal definition, monolith assessment, candidate selection, strangler fig pattern, data extraction, and validation.

Transforms a raw PRD, BRD, or ADO work item into a structured analysis document with extracted requirements, affected modules, test mappings, a pre-filled tech design, and a commit-level execution plan. Use when given a plain-English product requirements document that needs translating into actionable input for planning and implementation skills.

Creates a Task work item in Azure DevOps requesting access to an environment, assigned to TechOps. Use when you need access to a specific environment and can provide a reason.

Holistic audit of an existing microservice system against Newman's principles. Plan mode workflow covering boundaries, communication, workflows, deployability, observability, resiliency, security, scaling, and organizational alignment. Delivers prioritized findings and remediation recommendations.