cpliakas

cpliakas/claude-code-engineering-leaders

Agent Claude Code

Agile Coach peer agent for story quality review and retrospective facilitation. Peer of product-owner — consult for INVEST validation, AC quality checks, Definition of Done completeness, scope boundary review, and blameless retrospective facilitation. Use when a story draft needs coaching before backlog entry, to…

cpliakas/claude-code-engineering-leaders

Agent Claude Code

Strategic technical advisor who evaluates decisions against the product's long-term arc. Use when a change introduces a new pattern, touches data models or public contracts, spans multiple components, or involves choosing between approaches where the long-term trajectory matters. Also use when the user says…

cpliakas/claude-code-engineering-leaders

Agent Claude Code

DevOps strategy lead providing tool-agnostic guidance, maturity assessments, and operational doctrine. Sets the "what" and "why" of infrastructure decisions — does not write infrastructure code. Use for CI/CD pipeline design, deployment strategies, environment management, infrastructure architecture decisions…

cpliakas/claude-code-engineering-leaders

Agent Claude Code

SDLC meta-observer and systemic feedback agent. Scans PR review threads for deferred tech debt, monitors convention health, tracks architectural decision outcomes, analyzes code churn patterns, and generates SDLC health reports. Produces structured proposals for human approval — never enacts changes autonomously. Use…

cpliakas/claude-code-engineering-leaders

Agent Claude Code

Product owner and roadmap keeper. Use proactively for roadmap planning, work sequencing, phase prioritization, capturing strategic direction (vs. setting strategic direction), feature scoping, dependency analysis, or evaluating whether proposed work aligns with the project trajectory. Consult BEFORE starting…

cpliakas/claude-code-engineering-leaders

Agent Claude Code

QA strategy and test architecture lead. Use for test strategy, test type selection, coverage gap analysis, brittleness assessment, and risk-based test prioritization. Use when the user says "QA", "qa lead", "test strategy", "what should we test", "how should we test", "test types", "test architecture", "coverage…

cpliakas/claude-code-engineering-leaders

Agent Claude Code needs its repo

Tactical orchestrator during implementation and cross-domain convention registrar. Authors conventions in the tactical-implementation domain; routes convention authorship in other domains to the declared owner agent. Deconstructs stories into plans, routes to domain specialist agents, and reviews code for convention…

cpliakas/claude-code-engineering-leaders

Agent Claude Code

Strategic UX advisor who ensures the product experience is coherent across sprints, personas, and interaction surfaces. Use when evaluating whether a feature serves real user needs, when reviewing stories for user-observable outcomes, when a decision could create conflicting behavioral patterns, or when persona…

cpliakas/claude-code-engineering-leaders

Skill Claude Code

Register a specialist agent in the Tech Lead's routing model. Use when adding a new agent to the project that the Tech Lead should consult during implementation planning, incident analysis, or retrospectives.

cpliakas/claude-code-engineering-leaders

Skill Claude Code

Analyze code churn and thrash patterns in a git repository. Produces a structured report with hotspot detection, rework classification, temporal coupling, and severity-rated findings. Use for SDLC health checks, sprint retrospectives, or when recurring friction surfaces in code reviews.

cpliakas/claude-code-engineering-leaders

Skill Claude Code

Audit a single agent's project memory directory for hygiene issues. Use when you want to run agent memory hygiene checks, audit memory, inspect bloated memory, or detect state-like content, dead links, or oversized memory files. Complements /audit-routing-table, which audits the Tech Lead's specialist routing model.…

cpliakas/claude-code-engineering-leaders

Skill Claude Code

Audit the Tech Lead's specialist routing quality by reading the ## Routing Outcomes history and recommending narrowing actions for over-routed specialists. Complements /audit-routing-table (structural hygiene) with outcome-based data. Advisory only: never edits the routing table. Use when you want data-driven evidence…

cpliakas/claude-code-engineering-leaders

Skill Claude Code

Audit the Tech Lead's specialist routing model for orphan overrides, broken file pointers, redundant overrides, and thin agent descriptions. Reports findings for human review — does not auto-correct. Use after onboarding, after adding specialists, or when the Tech Lead appears to be missing consultation requests.

cpliakas/claude-code-engineering-leaders

Skill Claude Code

Conduct a blameless postmortem for an incident or failed/aborted release. Collects evidence and produces a structured postmortem document. Use when the user invokes /conduct-postmortem, says "postmortem", "incident retrospective", "release blocked", or describes a production incident or release failure that needs a…

cpliakas/claude-code-engineering-leaders

Skill Claude Code

Decompose a requirement into well-sequenced child items. Takes an epic and produces stories, or takes a story and produces subtasks.

cpliakas/claude-code-engineering-leaders

Skill Claude Code

Facilitate a structured retrospective from a sprint or body-of-work description. Follows the Derby-Larsen five-phase framework with blameless framing and SMART action items.

cpliakas/claude-code-engineering-leaders

Skill Claude Code

Use when configuring the Product Owner agent for a specific project. Gathers issue tracker details, backlog norms, and current roadmap state, then writes them to the Product Owner's project memory. Run /onboard first to establish shared project context. This is the reference implementation for per-agent onboarding…

cpliakas/claude-code-engineering-leaders

Skill Claude Code

Use when setting up the engineering-leaders plugin for the first time on a project, or when re-running onboarding to update shared project context. Gathers shared project context for all agents (one question at a time) and discovers specialist plugins for the Tech Lead to consult. Run this before per-agent onboarding…

cpliakas/claude-code-engineering-leaders

Skill Claude Code

Produce a test strategy for a feature, component, or story. Identifies the highest-impact tests by type and layer, flags brittleness risks, and recommends what NOT to test. Use for test planning, coverage gap analysis, or when deciding how to test a new feature.

cpliakas/claude-code-engineering-leaders

Skill Claude Code

Audit onboarded agent memory for drift against live project signals. Use when you want to re-onboard, detect onboarding drift, refresh project context, check for stale memory, or run an onboarding drift check. Compares each agent's project memory against cheaply derivable local signals (filesystem paths, specialist…

cpliakas/claude-code-engineering-leaders

Skill Claude Code

Score a story draft against INVEST criteria and eight agile coaching principles. Returns a structured report with pass/fail per dimension and specific rewrites or corrective actions for failures.

cpliakas/claude-code-engineering-leaders

Skill Claude Code

Convene the Product Owner, Chief Architect, and UX Strategist in parallel on a story draft and produce a consolidated readiness verdict. Use before /plan-implementation to surface scope, architectural, and persona-fit concerns while the story is still cheap to revise. Invoke with /refinement-review.

cpliakas/claude-code-engineering-leaders

Skill Claude Code

Produce a complete ADR (Architecture Decision Record) document in MADR format. Use when the user invokes /write-adr, the Chief Architect recommends recording a decision, or an architectural review surfaces decisions to document.

cpliakas/claude-code-engineering-leaders

Skill Claude Code

Scaffold a RIMGEN-validated bug report with reproduction steps, environment context, severity, and priority. Outputs named warnings for any missing or incomplete RIMGEN sections.