Plugin Claude Code
Plugin marketplace listing 3 plugins: dw-solo, dw-solo-setup, dw-solo-extras.
Plugin Claude Code
Plugin marketplace listing 3 plugins: dw-solo, dw-solo-setup, dw-solo-extras.
Hook Claude Code
Runs before the agent uses a tool for Bash, Read, Edit, Write, MultiEdit, NotebookEdit, Grep, Edit, Write, MultiEdit and NotebookEdit tool calls, executing bash-guard.sh, block-env-access.sh and guard-plugin-canon.sh via bash (3 commands). From dominikwozniak/dw-solo-skills.
Hook Claude Code
Runs after a tool call finishes for Write, Edit, MultiEdit and Write tool calls, executing lint-on-edit.sh and large-file-guard.sh via bash (2 commands). From dominikwozniak/dw-solo-skills.
Settings file Claude Code
Agent settings declaring 2 hook events (PreToolUse, PostToolUse).
Instructions file CodexOpenCode
AGENTS.md instructions for dominikwozniak/dw-solo-skills, covering dw-solo-skills — agent instructions, layout — and the one rule, the loop, commands and before you push.
Instructions file
Claude Code instructions for dominikwozniak/dw-solo-skills, a project described as: a persistent Solo Projects spec-to-ship workflow for Claude Code.
Agent
Artifacts are real work documents, committed with the code — not scratch.
Agent
The conventions themselves (commit format, trailer, branch naming) are in the root's ## Git conventions, because dw-git greps them there. This file is what goes wrong around them.
Agent
The root's layout rule — always edit skills/ /SKILL.md, never a plugins/… path — is absolute. This file is why it works and what it costs.
Agent
The commands live in the root's ## Commands; the gate is the scripts block of package.json. This file is how that tooling misbehaves.
Agent
Worktrees live at .claude/worktrees/ on branch ; scripts/runtime/worktree.sh owns create and remove. A claude -w session spells the branch worktree- instead, which is why dw-next strips that prefix before matching a change doc.
Plugin Claude Code
Off-loop skills for the solo lane — reached for beside the loop, not on the way from shape to ship. Four today: compact the live session into a handoff the next context window resumes from, mid-task; walk the backlog entry by entry, dropping what is stale and doing what is cheap; cut the AI tells out of the text a…
Plugin Claude Code
Project setup for the solo lane: scaffold the .ai/ state, guardrail hooks, settings, a tracked AGENTS.md under a declared size budget with the checker that holds it, and an optional pre-commit — all in one gated pass, then diagnose that it actually fires. Install alongside dw-solo; run once per repo.
Plugin Claude Code
The solo loop for repos only you read: interrogate a fuzzy idea into decisions, shape one CHANGE.md per change, build it task by task — in a worktree when changes run in parallel — check it fast, land the durable knowledge, ship the branch. One artifact per change, one thin quality pass. Pairs with dw-solo-setup…
Skill Claude CodeCodex
A fast, optional quality gate over the change in progress: the diff goes to an outside reviewer, or to a quick two-axis self-review — correct? fits this repo? — when the diff is trivial or no reviewer is installed, with findings at real file:line, verified, then fixed in-session after approval.
Skill Claude CodeCodex
Read-only environment diagnostic for a solo-lane repo: whether the tools the hooks and skills assume are installed and whether the repo's guardrails will really fire, each gap reported with a copy-paste fix.
Skill Claude CodeCodex
One skill for every git operation in this project — commit and save what you have, stage, push, open a pull request, sync with main, rebase, branch, park edits in a stash — applying the repo's own ## Git conventions from AGENTS.md instead of generic defaults.
Skill Claude CodeCodex
Audit code just written for excess the gate cannot see: a helper reinvented beside its canonical home, a wrapper with one caller, config for a case that never occurs, a shape the neighbouring files build differently, code this change stranded. One table, and every row is a deletion or a reuse. Explicit-invoke only.
Skill Claude CodeCodex
Interview the user to sharpen a vague, half-formed idea into decisions before any of it gets built — one question at a time, at most five, hardest-first, each with a recommended answer, so the thinking is finished first.
Skill Claude CodeCodex
Compact the live session into .ai/work/ /HANDOFF.md — how far into the current task you are, what is applied but uncommitted, and which dead ends are already ruled out, so the next context window resumes mid-task. Explicit-invoke only.
Skill Claude CodeCodex
Scaffold a private/solo repo for the solo loop — .ai/ (work / backlog / archive), docs/decisions/, CONTEXT.md, a tracked AGENTS.md with a Task Router and a size budget, the guardrail hooks, settings with a derived allow-list, and an optional pre-commit. Explicit-invoke only.
Skill Claude CodeCodex
Close out a finished change and decide whether it is ready to merge: one thin verdict over the diff, then on your go promote the durable residue to docs/decisions/, CONTEXT.md, ## Gotchas and the backlog, archive the change doc, and push the branch and open its PR.
Skill Claude CodeCodex
The solo lane's build step and its resume point in one skill: report where the active .ai/work/ /CHANGE.md stands — read from disk, so it survives a /clear — then build every remaining unticked task, one commit each.
Skill Claude CodeCodex
Walk .ai/backlog/ and decide every entry out loud: drop what is stale or already done, do what is cheap and unblocked, bundle what ships with a cousin, leave the rest. Explicit-invoke only.