Skill Claude CodeCodex
Decision guide for delegating to caveman-style subagents. Tells the main thread WHEN to spawn cavecrew-investigator (locate code), cavecrew-builder (1-2 file edit), or cavecrew-reviewer (diff review) instead of doing the work inline or using vanilla Explore. Subagent output is caveman-compressed so the tool-result…
Skill Claude CodeCodex
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.…
Skill Claude CodeCodex
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…
Skill Claude CodeCodex
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".
Skill Claude CodeCodex
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.
Skill Claude CodeCodex
Show real token usage and estimated savings for the current session. Reads directly from the Claude Code session log — no AI estimation. Triggers on /caveman-stats. Output is injected by the mode-tracker hook; the model itself does not compute the numbers.
Skill Claude CodeCodex
Ultra-compressed communication mode. Cuts output tokens 65% (measured) 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"…
Skill Claude CodeCodex
Close out a bead started with /work-on — verify it's safe, clean up the worktree, free the tab and the session name.
Skill Claude CodeCodex
Grant b4mad-release-agent what it needs to run semantic-release on a Forgejo repo, install just release/just preflight into that repo's justfile and a short how-to into its AGENTS.md, then print the oc command that runs the release. Use when asked to "enable semantic-release for ", to onboard a repo to the release…
Skill Claude CodeCodex
Start focused work on a bead — rename the tab and the Claude session, create the worktree, and load the issue.
Skill Claude CodeCodex
Guidance for Worktrunk (the wt CLI) — git worktree management, hooks, and config. Load when editing .config/wt.toml or /.config/worktrunk/config.toml; adding, modifying, or debugging hooks (post-merge, post-start, pre-commit, pre-merge, post-switch, etc.); configuring commit message generation or command aliases; or…
Skill Claude CodeCodex
Create a new worktrunk worktree (optionally in another repo) and switch this session's working directory into it. Use when launching a session that should work in its own worktree.
Agent Claude Code
Adversarial reviewer of Spellkave research notes, ADRs, and OpenSpec proposals. Surfaces only load-bearing critiques — things that would actually break the system in production or modding. Use when a proposal feels too clean and you need a hard "what would break?" pass.
Agent Claude Code
Implements a bounded scope assigned by the lead during team-mode apply. The scope may be one task or a short contiguous task range. Reads the assigned artifacts and repository guidance, edits only what the scope requires, reports per-task status, and leaves workflow state to the lead.
Agent Claude Code
Defends a target Spellkave proposal against adversarial critique. Pairs with the devils-advocate teammate in a debate team. Use when you want a structured back-and-forth instead of single-agent simulation.
Agent Claude Code
Autonomous owner of one feasibility spike inside sandbox/. Runs the entire loop — scaffold, build, deploy if module, test, measure, report — without user intervention. Ignores the sandbox's "user runs every CLI" interactive-learning rule (that rule exists for the human's benefit; an agent has nothing to learn by…
Agent Claude Code
Independently verifies an implemented scope during team-mode apply. The scope may be one task or a short contiguous task range. Reads artifacts, repository guidance, diff, and tests without relying on implementer reasoning. Reports pass/fail/concerns per task and does not edit files.
Command Claude Code
Review last commit and branch changes.
Command Claude Code
Multi-agent PR code review for Forgejo repositories.
Command Claude Code
Stage all changes, run pre-commit hooks, fix failures, and create conventional commit with emoji.
Command Claude Code
Implement tasks from an OpenSpec change (Experimental).
Command Claude Code
Archive a completed change in the experimental workflow.
Command Claude Code
Enter explore mode - think through ideas, investigate problems, clarify requirements.
Command Claude Code
Propose a new change - create it and generate all artifacts in one step.