Copy this prompt into the AI assistant you're using.
Follow https://skillsmp.com/skill-install/prompt.md to install Agent Skills from https://github.com/LumiroCode/copilot-instructions.Loading the repository README…
Showing 37 of 37 collected skills.
Expert-level software engineering agent. Deliver production-ready, maintainable code. Execute systematically, interactively, and specification-driven. Document comprehensively. Gather evidence, then involve the user for approvals and business-direction…
Implement a single user-indicated phase from a whole plan.
Research, align business decisions with the user, and produce a locked phased implementation plan.
Review code changes, suggest improvements, and ensure code quality.
Orchestrates bugfixing with help of specialised subagents.
Orchestrates complex features development.
Codex-native counterpart of `/home/projekty/copilot-instructions/prompts/Solve_next_todo_item.prompt.md`.
Testing specialist for raw unit-failure capture and Playwright manual browser validation with actionable QA insights.
Implementation directive with consolidation of most important rules.
Researches and outlines multi-step plans - adjusted by me
Strict architecture guard for design/code review against docs/architecture.md, with hard failures for missing architecture source, strong separation-of-concerns checks, blackbox modularity enforcement, and overcomplication detection.
Produces final-goal Definition of Done and acceptance criteria from current app state, user impact, and UX evidence. Works interactively with the user through iterative refinement. This agent must not investigate solutions, implementation steps, or technical…
Challenge assumptions and encourage critical thinking to ensure the best possible solution and outcomes.
Expert at debugging web applications, analyzing errors, and identifying root causes
Perform janitorial tasks on any codebase including cleanup, simplification, and tech debt remediation.
Use when you need a painstaking, robotic analysis of ONE question from every relevant angle and then a focused, evidence-based answer. Triggers: deep one-question analysis, exhaustive single-question assessment, angle-by-angle reasoning, highly thorough…
Orchestrates post-coding finalization as an assessment-only pipeline: deep review, sanity checks, testing, and verification with triage-ready findings.
Selecting resolution for a problem base on RCA
Business-value and product-alignment reviewer. Judges changes by user-flow soundness, business value, product vision fit, and whether the result strengthens or weakens the product direction.
Specializes in reproducing production issues from user reports, log analysis, and performance metrics. Produces structured triage reports.
RCA analyzis of a problem
Verify implementation and tests against acceptance criteria using objective evidence and a strict pass/fail report.
Generate a unique conversation ID for orchestrator agents to namespace `/memories/session/*` artifacts and prevent collisions across parallel chats/subagent delegations. Produces an 8-character hex hash from random bytes.
Perform a sanity check (Why-WhatFor analysis) on your work before finalizing an approach or code change.
Perform a sanity check (Why-WhatFor analysis) on a decision before proceeding with it.
Verify layering violations, dependency direction, event-driven directionality-first data flow (directional, coherent, bounded), MerlinX ACL boundary integrity, and separation of concerns against docs/architecture.md.
Verify layering violations, dependency direction, event-driven directionality-first data flow (directional, coherent, bounded), MerlinX ACL boundary integrity, and separation of concerns against docs/architecture.md.
Generate a unique conversation ID for orchestrator agents to namespace `.agents/session/*` artifacts and prevent collisions across parallel chats/subagent delegations. Produces an 8-character hex hash from random bytes.
Run core PHP tests via tests/run_all.sh and return only failed test files with exact error/throw/fail lines. Use when user asks for failure-only raw output with no narrative summary.
Audit a conversation for undocumented decisions that meet logging criteria (durable, cross-cutting, non-obvious, normative).
Run the post-coding finalization pipeline: deep review, sanity checks/tests, verification, and triage-ready findings (no fixes).
Validate input sanitization, auth/authz enforcement, secrets exposure, and OWASP Top 10 patterns in changed code.
Analyze test files for meaningful assertions that process actual data; verify changed code is exercised by tests.
Run core PHP tests via tests/run_all.sh and return only failed test files with exact error/throw/fail lines. Use when user asks for failure-only raw output with no narrative summary.
Audit a conversation for undocumented decisions that meet logging criteria (durable, cross-cutting, non-obvious, normative).
Validate input sanitization, auth/authz enforcement, secrets exposure, and OWASP Top 10 patterns in changed code.
Analyze test files for meaningful assertions that process actual data; verify changed code is exercised by tests.
The repository README could not be loaded. You can still explore its Skills below.