dotfiles

Use when executing a multi-step plan and each implementation step should be delegated to subagents and independently reviewed.

Use when reviewing local changes — the working-copy diff, a branch, a commit, or a GitHub PR by number — with fresh reviewer subagents that return structured findings.

Use when a test failure, regression, exception, hang, wrong result, or unexpected behavior needs diagnosis — gathers evidence, traces the relevant path, and verifies the cause before recommending a fix.

Resolve jj (Jujutsu) conflicts. Use when jj log/status shows conflicted revisions, a rebase/squash/abandon reports 'new conflicts appeared', or files contain jj conflict markers.

Use when you want to work in an isolated jj working copy — parallel task, experimental scratch, subagent with its own tree. jj's equivalent of git worktrees: creating a workspace, working inside it from anywhere, and cleaning up without losing history.

Find things to improve in the current project

Build a mental model of code changes — guided reading order, conceptual grouping, and context.

Use when the user asks to be exhaustive, thorough, comprehensive, or rigorous, wants a finding drilled into or adversarially verified, or when missing something would be costly — a deliberate deep-work mode that trades speed for rigor.

Use when writing skills, CLAUDE.md files, agent prompts, or shell snippets that touch environment variables, API credentials, file creation, or git operations. Covers keeping secrets out of context, safe shell patterns, and credential exposure.

Use when verifying that a skill works before relying on it. Pressure-tests skills with subagents — runs the scenario without the skill, compares to with the skill, iterates until the behavior is reliable.

Use when writing instructions that guide an agent: skills, CLAUDE.md/AGENTS.md files, system prompts, agent prompts. Covers token efficiency, discovery, compliance, and the small repertoire of patterns that earn their keep.

Use when creating or updating CLAUDE.md / AGENTS.md files for a project or a subdirectory. Covers the split between top-level (how to work here) and domain-level (why this exists, what it promises), and the freshness date convention.

Use when creating, editing, or reviewing a skill before it ships. Covers when to write a skill, file layout, the SKILL.md shape, and how to verify it before relying on it.

Use when writing skills, CLAUDE.md files, agent prompts, or shell snippets that touch environment variables, API credentials, file creation, or git operations. Covers keeping secrets out of context, safe shell patterns, and credential exposure.

Use when verifying that a skill works before relying on it. Pressure-tests skills with subagents — runs the scenario without the skill, compares to with the skill, iterates until the behavior is reliable.

Use when writing instructions that guide an agent: skills, CLAUDE.md/AGENTS.md files, system prompts, agent prompts. Covers token efficiency, discovery, compliance, and the small repertoire of patterns that earn their keep.

Use when creating or updating CLAUDE.md / AGENTS.md files for a project or a subdirectory. Covers the split between top-level (how to work here) and domain-level (why this exists, what it promises), and the freshness date convention.

Use when creating, editing, or reviewing a skill before it ships. Covers when to write a skill, file layout, the SKILL.md shape, and how to verify it before relying on it.

Relentlessly interview the user about a plan or design.

Adversarial review/drafting of persuasive technical prose.

Use when committing, rebasing, inspecting history, or fixing repo state with jj (Jujutsu) — including colocated .git/.jj repos, detached-HEAD confusion, and stale-workspace errors (read this before running `jj workspace update-stale`; it can silently discard uncommitted changes). Flags interactive commands that hang agents.