meganemura/*

How the tools use each other

An arrow points from a tool to a tool it depends on. The edges come from each package's runtime and development dependencies on npm and RubyGems, the most used tool sits in the center, and a two-headed colored line marks two tools that use each other.

Each tool was also run on the other repositories, and what got in the way was fixed in the tool itself: how the tools fixed each other.

Command-line tools on npm

Every version, description, and install command below is read from the npm registry, not typed by hand.

• archstrict v0.4.0

  TypeScript module boundary checker: default-private module surfaces, layers, cycles, and a todo freeze that can only shrink.

  npm install archstrict

  Uses: depug, nukadoko, stryker-agent-reporter

  Source on GitHub

• autospawn v0.3.0

  Attach-or-spawn launcher for stdio programs, such as MCP servers, so a secret-resolving wrapper starts once and every client shares it.

  npm install autospawn

  Uses: archstrict, finderscope, stryker-agent-reporter

  Source on GitHub

• depug v0.2.0

  Measure runtime values in TypeScript test runs instead of guessing them from source.

  npm install @meganemura/depug

  Uses: archstrict, nukadoko

  Source on GitHub

• finderscope v0.4.0

  Turn V8 CPU and heap profiles into short, ranked reports with runnable next commands for coding agents.

  npm install finderscope

  Uses: archstrict, nukadoko, stryker-agent-reporter

  Source on GitHub

• firstlight v0.3.0

  Run lint, typecheck, and tests in parallel for a coding agent, and return at the first failure.

  npm install firstlight

  Uses: archstrict, depug, nukadoko, stryker-agent-reporter

  Source on GitHub

• nukadoko v0.17.0

  Typed step contracts for Gherkin, so a scenario can be told wrong before it runs. Every execution leaves a step record the tool measured rather than the agent reported, and every run a Markdown summary a pull request carries.

  npm install nukadoko

  Uses: archstrict, depug

  Source on GitHub

• polylinedb v0.5.2

  Polyline database: project memory and task tracking for local tools and cloud agents

  npm install polylinedb

  Uses: archstrict, solarsql, stryker-agent-reporter

  Source on GitHub

• portful v0.1.0

  Keep one fixed port pointed at whichever git worktree is active, and switch it without restarting the client

  npm install portful

  Source on GitHub

• runtag v0.1.0

  Record one child process under XDG so an agent can observe pid and exit later.

  npm install runtag

  Source on GitHub

• solarsql v0.8.0

  Typed SQL for SQLite on Cloudflare D1 and Durable Objects: types from the real engine at build time, a command as one transaction, failures as values.

  npm install solarsql

  Uses: archstrict, nukadoko, stryker-agent-reporter

  Source on GitHub

• spacequery v0.4.0

  Named queries over the state of one developer's machine: which coding agents run where, which repositories are dirty, which worktrees and sessions sit idle.

  npm install spacequery

  Uses: archstrict, nukadoko, solarsql

  Source on GitHub

• stow v0.1.1

  Personal skills for Git checkouts with local host links

  npm install @meganemura/stow

  Source on GitHub

• stryker-agent-reporter v0.3.0

  A Stryker reporter that writes JSON Lines a coding agent can act on directly

  npm install stryker-agent-reporter

  Uses: archstrict, nukadoko

  Source on GitHub

Ruby gems

Every version, description, and install command below is read from RubyGems.

• archspec_diagram v0.1.0

  Draw an Archspec.rb as a Mermaid flowchart

  gem install archspec_diagram

  Source on GitHub

• bulldogger v0.3.1

  Ruby execution evidence for coding agents: failure snapshots captured at the moment a test raises, written as files an agent can read and query.

  gem install bulldogger

  Uses: hegel-ruby

  Source on GitHub

• hegel-ruby v0.3.1

  An unofficial Ruby implementation of the Hegel protocol. Hegel is a property-based testing framework based on Hypothesis. This gem drives libhegel, the same native engine that backs the Rust, Go, TypeScript, Java, OCaml, and C++ implementations.

  gem install hegeltest

  Source on GitHub

Agent plugins and skills

These have no package registry. Each description, version, license, and skill name below is read from the repository’s own manifest, skill files, README, and LICENSE at the latest commit of its default branch, and each install command is checked against its README. Each page says which agents the repository supports.

• headsign

  Skills that help a coding agent create and improve a project's way of working.

  gh skill install meganemura/headsign --all --agent codex --scope user

  Source on GitHub

• mutation-testing-skills

  Claude Code skills for mutation testing. Each skill covers one tool and walks an agent through setup, running, reading results, and fixing surviving mutants.

  npx skills add meganemura/mutation-testing-skills --skill stryker -g -a claude-code

  Source on GitHub

• pstack-portable v0.4.5

  Run unmodified upstream pstack in Claude Code, Codex, and Cursor through a thin execution wrapper.

  claude plugin marketplace add meganemura/pstack-portableclaude plugin install pstack-portable@pstack-portableclaude plugin install pstack@pstack-portable

  Source on GitHub

• steerhook

  Rules that block or warn before a tool runs and tell Claude the alternative. Rules live in ~/.claude/steerhook/ and apply in every project.

  /plugin marketplace add meganemura/steerhook/plugin install steerhook@steerhook

  Source on GitHub

Claude Code mods

Each of these is a Claude Code plugin, which its README calls a Claude Mod, that draws a pane beside the transcript. They work only in Claude Code, and each one needs CLAUDE_CODE_ENABLE_FUNCTION_HOOKS=1. Their facts are read from each repository in the same way as for the plugins and skills above.

• buffer-pane v0.2.0

  A pane beside the transcript that holds the next things to tell the agent: write blocks of text while a turn runs, then press [+] to put one block into the prompt box.

  claude plugin marketplace add meganemura/buffer-paneclaude plugin install buffer-pane@buffer-pane

  Source on GitHub

• draft-pane v0.5.0

  A pane beside the transcript that shows the model's newest open draft and sends your comments on it, span by span, as one prompt.

  claude plugin marketplace add meganemura/draft-paneclaude plugin install draft-pane@draft-pane

  Source on GitHub

• grilling-pane v0.5.0

  A pane beside the transcript that shows the model's open interview questions and submits every pick as one prompt.

  claude plugin marketplace add meganemura/grilling-paneclaude plugin install grilling-pane@grilling-pane

  Source on GitHub

• jira-ticket-pane v0.4.0

  A pane beside the transcript that fetches one Jira issue by key and lets you attach it to your next prompt.

  claude plugin marketplace add meganemura/jira-ticket-paneclaude plugin install jira-ticket-pane@jira-ticket-pane

  Source on GitHub

• pull-request-pane v0.4.0

  A pane beside the transcript with the GitHub pull requests and issues related to the current session: drag a title or description to leave a comment on it, add as many as you like, then Submit sends them all as one prompt. Watch each pull request's checks, review decision and mergeability update on a timer.

  claude plugin marketplace add meganemura/pull-request-paneclaude plugin install pull-request-pane@pull-request-pane

  Source on GitHub

Sentences that run

Some sentences on the tool pages carry a small mark such as “checked with archstrict”. That sentence is tied to a command in this repository. Before each deploy, npm run claims runs the command and compares its exit code and output with what the sentence says. If the tool changes and the sentence stops being true, the check fails and names the sentence.