Skills
Provision Instant Temporary Postgres Databases
Provision temporary, no-signup Postgres databases for prototyping and testing using the pg.new service. It supports CLI, SDK, and REST API for rapid environment setup.
Zed Brand Voice Writer
An agentic skill for generating and reviewing technical copy that adheres to Zed's specific brand voice and principles. It employs a multi-pass drafting and scoring system to ensure technical grounding and the avoidance of marketing tropes.
AI Writing Pattern Remover
This skill identifies and removes common linguistic patterns that characterise LLM outputs. It rewrites text to inject personality and vary sentence structure, helping to optimise copy for a more natural, human-written feel.
Zed Brand Voice Writer
An agentic skill for generating and reviewing technical copy that adheres to Zed's specific brand voice and principles. It employs a multi-pass system to ensure technical grounding and eliminate marketing tropes.
Automated Change Proposal and Artifact Generation
This skill automates the creation of comprehensive change proposals by generating design documents, specifications, and implementation tasks in a single workflow. It leverages the openspec CLI to scaffold directories and iterate through req…
OpenSpec Exploration and Discovery Skill
An agentic skill for deep problem investigation, requirement clarification, and architectural exploration without implementing code. It utilises codebase analysis and ASCII visualisation to assist developers in evaluating options and docume…
Archive completed openspec changes
Finalises and archives completed changes within the experimental workflow by verifying artifact completion, task status, and synchronising delta specs. It automates the movement of change directories to a dated archive folder after ensuring…
Implementing tasks from OpenSpec changes
This skill automates the implementation of tasks defined within an OpenSpec change by iteratively processing instructions and updating task progress.
Human Pages: AI-to-Human Task Discovery
An AI-to-human discovery layer for finding and hiring real-world service providers for tasks such as photography, research, or delivery. It enables agents to search for professionals, manage job offers, and process payments via crypto or fi…
Agentic Plan and Implementation Review Protocol
Automates the validation of project plans and implementations by detecting review context and orchestrating specialist sub-agents. It performs requirements traceability, identifies plan deviations, and verifies changes through git diffs and…
Claude Skill Definition Auditor
Audits .claude/skills/ definitions for frontmatter correctness, trigger clarity, and progressive disclosure hygiene. It identifies redundancy and gap candidates, automatically generating board tasks for all identified issues.
Claude Configuration Security Auditor
Performs a security audit of the .claude/ configuration layer to identify leaked secrets, overly broad permissions, and vulnerable hook scripts. It evaluates MCP server tool exposure and dangerous agent instructions, automatically generatin…
Core Instruction Layer Review Protocol
Audits CLAUDE.md, rule files, and settings hooks for size, staleness, and cross-reference integrity. It identifies redundancies and alignment issues, automatically generating board tasks for any discovered discrepancies.
Claude Code Memory Audit Skill
Audits the Claude Code auto-memory directory to ensure index integrity, validate frontmatter, and identify stale or duplicate entries. It automatically generates board tasks for findings such as uncaptured patterns or rule-promotion candida…
Automated Documentation Audit and Maintenance
Audits documentation trees for broken links, stale content, and structural inconsistencies, automatically generating project board tasks for every identified issue.
Vibe Board Audit and Remediation Skill
Audits Firebase Firestore task boards for stale, orphaned, or duplicate tasks and identifies referential integrity issues. Automatically generates remediation tasks to maintain board hygiene and process compliance.
Full System Audit Orchestrator
This orchestrator executes a sequential suite of review skills to perform comprehensive system audits across agents, skills, rules, and documentation. It aggregates findings into a single board parent task and generates a prioritised remedi…
Claude Agent Configuration Auditor
Automates the audit of .claude/agents/ markdown files to ensure accuracy, model assignment efficiency, and compliance with Anthropic best practices. The skill identifies stale references, merge/split candidates, and gap analysis opportuniti…
Agentic Implementation Planning Protocol
A protocol for generating structured implementation plans by decomposing tasks into actionable board items and assigning them to specialised agents. It ensures rigorous execution through detailed acceptance criteria and automated task creat…
Autonomous Full-Cycle Task Orchestrator
Automates the complete RIPER cycle by orchestrating research, planning, execution, and review phases autonomously. It includes built-in safety guards for managing scope escalation, line change limits, and critical review failures.
ve-kit project bootstrap protocol
An interactive protocol for setting up or upgrading projects within the ve-kit framework. It automates directory structure creation, rule generation, hook installation, and MCP server configuration.
Automated subproject release workflows
Automates the building and publishing of subprojects to PyPI and Docker Hub via GitHub Actions. It manages the release lifecycle for various Memgraph-related packages and integrations.
Agent Context Graph for Codex Capture
This skill facilitates the setup and verification of Codex activity capture into Memgraph, managing runtime hook wiring and providing diagnostic capabilities for event-driven graph updates.
Memgraph Agent Context Graph for Claude Code
This skill enables the capture of Claude Code activity into a Memgraph graph database by installing runtime hooks and event connectors. It facilitates the monitoring and verification of event flows from the Claude Code runtime to the graph.