pjmagee

Generate tests for a completed phase based on UAT criteria and implementation

Generate an AI-SPEC.md design contract for phases that involve building AI systems.

Autonomous audit-to-fix pipeline — find issues, classify, fix, test, commit

Audit milestone completion against original intent before archiving

Cross-phase audit of all outstanding UAT and verification items

Run all remaining phases autonomously — discuss→plan→execute per phase

Capture ideas, tasks, notes, and seeds to their destination

Archive accumulated phase directories from completed milestones

Author or update a numbered ADR under `eng/adr/`, or update the header/status of a historical design narrative migrated to `specs/[NNN]-[slug]/spec.md`. Enforces the shared schema — filename pattern, H1, required header fields, the fixed `Status` enumeration, ISO-8601 dates, the canonical `Related:` link list. Trigger when the user says "write an ADR", "update the status of …", "supersede …", or is editing a file under `eng/adr/`. For NEW feature design narratives, use `/speckit-specify` (spec-kit) instead — `eng/design/` was retired in constitution v2.0.0 (2026-05-24).

Manages MongoDB Atlas Stream Processing (ASP) workflows. Handles workspace provisioning, data source/sink connections, processor lifecycle operations, debugging diagnostics, and tier sizing. Supports Kafka, Atlas clusters, S3, HTTPS, and Lambda integrations for streaming data workloads and event processing. NOT for general MongoDB queries or Atlas cluster management. Requires MongoDB MCP Server with Atlas API credentials.

Ultra-compressed commit message generator. Cuts noise from commit messages while preserving intent and reasoning. Conventional Commits format. Subject ≤50 chars, body only when "why" isn't obvious. Use when user says "write a commit", "commit message", "generate commit", "/commit", or invokes /caveman-commit. Auto-triggers when staging changes.

Compress natural language memory files (CLAUDE.md, todos, preferences) into caveman format to save input tokens. Preserves all technical substance, code, URLs, and structure. Compressed version overwrites the original file. Human-readable backup saved as FILE.original.md. Trigger: /caveman:compress <filepath> or "compress memory file"

Quick-reference card for all caveman modes, skills, and commands. One-shot display, not a persistent mode. Trigger: /caveman-help, "caveman help", "what caveman commands", "how do I use caveman".

Ultra-compressed code review comments. Cuts noise from PR feedback while preserving the actionable signal. Each comment is one line: location, problem, fix. Use when user says "review this PR", "code review", "review the diff", "/review", or invokes /caveman-review. Auto-triggers when reviewing pull requests.

Ultra-compressed communication mode. Cuts token usage ~75% by speaking like caveman while keeping full technical accuracy. Supports intensity levels: lite, full (default), ultra, wenyan-lite, wenyan-full, wenyan-ultra. Use when user says "caveman mode", "talk like caveman", "use caveman", "less tokens", "be brief", or invokes /caveman. Also auto-triggers when token efficiency is requested.

This skill provides guidance for creating sophisticated, interactive data visualisations using d3.js.

Cut and publish a release of this mod end-to-end: pre-release gates (merged PRs, CI, build, offline validation, the user's in-game checklist, SFSE-log health check), the three-place version bump, CHANGELOG via the changelog skill, tag -> CI (GitHub release + Nexus auto-upload), and the manual Nexus page steps. USE WHEN: user asks to cut/ship/tag/publish a release, "prep v1.6.0", "release this", "push to Nexus", or asks what's left before a release. COVERS: release gates, version-field sync (xmake.lua + Plugin.h + the CompatibleVersions duty), tag flow, Nexus group-id/paste steps, git signing gotcha, post-release verification.

Use the xAI **grok** CLI as a SEPARATE-USAGE subagent (its own grok.com billing, NOT Claude usage) to (a) run independent adversarial CODE REVIEWS of branches/diffs/PRs, and (b) offload background C++/Papyrus/RE/docs work in an isolated git worktree. Grok runs headless/non-interactively (`grok -p` / `--prompt-file`, JSON output) and already reads this repo's CLAUDE.md + skills. Use whenever you want a second opinion that doesn't spend Claude tokens, a review gate before committing risky native/offset/Papyrus changes, or extra parallel agents. Full CLI in references/cli.md.

End-to-end guidance for modding and reverse engineering Bethesda's Starfield: SFSE plugin development with CommonLibSF, Papyrus scripting and decompilation (Champollion), binary analysis in Ghidra (GUI and `analyzeHeadless` CLI for scripted/agent-driven RE), address-library offset discovery across game versions, and packaging/distribution via FOMOD, Wrye Bash, and Nexus Mods. USE WHEN: user works on Starfield mods (SFSE plugins, Papyrus scripts, ESM/ESP plugins), reverse engineers Starfield.exe, hunts for function offsets after a game update, decompiles .pex scripts, drives Ghidra in headless mode for automated analysis, or packages/distributes a mod. COVERS: SFSE plugin anatomy, address library IDs, CommonLibSF RE/ types, Papyrus <-> native boundaries, Ghidra signature hunting (interactive and headless `analyzeHeadless` CLI workflows), AOB/sig updating workflow, FOMOD structure, Wrye Bash tagging, community-patch conventions.

Maintain CHANGELOG.md for this mod with honest accounting: diff git tags, filter tooling/docs noise, and separate genuine player/runtime behavior changes from refactors and recompiles. USE WHEN: user asks "what changed between vX and vY", to update/write the changelog, to cut or document a release, to draft Nexus release notes, or asks for an honest summary of a version bump. COVERS: tag-to-tag diffing, noise filtering, Keep a Changelog formatting, the project's behavior-vs-refactor convention, Nexus-facing notes.

Debug LLM applications using the Phoenix CLI. Fetch traces, analyze errors, review experiments, inspect datasets, and query the GraphQL API. Use when debugging AI/LLM applications, analyzing trace data, working with Phoenix observability, or investigating LLM performance issues.

Build and run evaluators for AI/LLM applications using Phoenix.

OpenInference semantic conventions and instrumentation for Phoenix AI observability. Use when implementing LLM tracing, creating custom spans, or deploying to production.