claude

house-rules

Enforces aj's global CLAUDE.md rules as hooks: no hidden background work, no unasked git mutations, no unchecked destructive commands.

1 on GitHubby ajv2.80.0updated 3 days ago
/

Plugin house-rules from marketplace aj-house-rules in Ajw2003/AjsClaudeCodeTools. Commands read from the repo's manifest — 3 days ago.

This plugin runs code on your machine — its hooks fire automatically on events like tool use. The exact commands are listed below; read them before installing.

What it installs

Skills (2)

  • plain-docs

    Write a plain-English copy of a tier-4 system doc (or another eligible tier), for a

  • project-docs

    The six-tier documentation structure every repo uses - the tier spec, the per-tier templates, the conventions that keep it honest, and scaffolding for a repo that has none of it yet. Use when creating a repo's documentation, restructuring docs that grew without a shape, auditing docs against the cod

Slash commands (6)

  • /house-rules:docref

    Check, and optionally repair, the doc-ref pointers the archivist leaves in code

  • /house-rules:doctor

    Diagnose why house-rules hooks aren't running, and propose the fix for this machine

  • /house-rules:harvest-scan

    Scan a project, project-wide, for long-form comments the harvest hook would flag

  • /house-rules:labels

    Set up the three house labels on this repo, or on a repo you name, creating only the missing ones

  • /house-rules:mode

    Switch the design loop or the builder helpers on or off for this repo, or show which are off

  • /house-rules:plain-docs

    Check the plain-English doc copies under docs/plain/ against the plain-docs rules

Subagents (5)

  • archivist

    Moves long-form comment blocks out of source files into the tier-4 system documents that own them, leaving a one-line pointer behind. Use PROACTIVELY the moment the comment-harvest hook reports blocks in a file, or when asked to port comments, essays or design notes into docs - delegate here without

  • builder

    Implements one issue's chunk of work (three steps or fewer) from a plan that is already decided. Use PROACTIVELY for implementation once a plan is settled; the parent writes the job-specific steps into the spawn prompt.

  • editor

    Rewrites a page or a long reply so it reads as proper English in a human voice and leaves nothing for aj to work out. Use PROACTIVELY before publishing any page under docs/plans/, and, when HOUSE_RULES_EDITOR=on, before sending a reply of 160 words or more - the end-of-reply check reminds you when i

  • reviewer

    Adversarial check of a finished diff. Use only when a second opinion is asked for, or after a builder has failed twice. Reports findings only.

  • scout

    Read-only lookup. Use PROACTIVELY for any "where is X", "what does Y do" or "list the Z" question instead of searching on the main thread. Finds, lists, reads and summarises code or docs and reports back; never edits and never runs commands.

Hooks (30)

  • SessionStart

    sh "${CLAUDE_PLUGIN_ROOT}/scripts/run.sh" inject

  • SessionStart

    sh "${CLAUDE_PLUGIN_ROOT}/scripts/run.sh" showrule

  • SessionStart

    sh "${CLAUDE_PLUGIN_ROOT}/scripts/run.sh" issuelist

  • SessionStart

    sh "${CLAUDE_PLUGIN_ROOT}/scripts/run.sh" profile

  • SessionStart

    sh "${CLAUDE_PLUGIN_ROOT}/scripts/run.sh" standards

  • SessionStart

    sh "${CLAUDE_PLUGIN_ROOT}/scripts/run.sh" docstiers

  • SessionStart

    sh "${CLAUDE_PLUGIN_ROOT}/scripts/run.sh" diskreport

  • SessionStart

    sh "${CLAUDE_PLUGIN_ROOT}/scripts/run.sh" versioncheck

  • UserPromptSubmit

    sh "${CLAUDE_PLUGIN_ROOT}/scripts/run.sh" scope

  • UserPromptSubmit

    sh "${CLAUDE_PLUGIN_ROOT}/scripts/run.sh" userpromptaudit

  • UserPromptSubmit

    sh "${CLAUDE_PLUGIN_ROOT}/scripts/run.sh" worktreesweep

  • PreToolUse

    sh "${CLAUDE_PLUGIN_ROOT}/scripts/run.sh" guard

  • PreToolUse

    sh "${CLAUDE_PLUGIN_ROOT}/scripts/run.sh" guardwrite

  • PreToolUse

    sh "${CLAUDE_PLUGIN_ROOT}/scripts/run.sh" guardgithub

  • PreToolUse

    sh "${CLAUDE_PLUGIN_ROOT}/scripts/run.sh" agentcap

  • PreToolUse

    sh "${CLAUDE_PLUGIN_ROOT}/scripts/run.sh" commitgate

  • PreToolUse

    sh "${CLAUDE_PLUGIN_ROOT}/scripts/run.sh" pagecheck

  • Stop

    sh "${CLAUDE_PLUGIN_ROOT}/scripts/run.sh" handover

  • PostToolUse

    sh "${CLAUDE_PLUGIN_ROOT}/scripts/run.sh" artifact

  • PostToolUse

    sh "${CLAUDE_PLUGIN_ROOT}/scripts/run.sh" branchnudge

  • PostToolUse

    sh "${CLAUDE_PLUGIN_ROOT}/scripts/run.sh" creditcheck

  • PostToolUse

    sh "${CLAUDE_PLUGIN_ROOT}/scripts/run.sh" closablemerge

  • PostToolUse

    sh "${CLAUDE_PLUGIN_ROOT}/scripts/run.sh" runnable

  • PostToolUse

    sh "${CLAUDE_PLUGIN_ROOT}/scripts/run.sh" harvest

  • PostToolUse

    sh "${CLAUDE_PLUGIN_ROOT}/scripts/run.sh" delegate

  • PostToolUse

    sh "${CLAUDE_PLUGIN_ROOT}/scripts/run.sh" audit

  • PostToolUse

    sh "${CLAUDE_PLUGIN_ROOT}/scripts/run.sh" rclink

  • PostToolUse

    sh "${CLAUDE_PLUGIN_ROOT}/scripts/run.sh" autosave

  • PostToolUse

    sh "${CLAUDE_PLUGIN_ROOT}/scripts/run.sh" promptran

  • PostToolUseFailure

    sh "${CLAUDE_PLUGIN_ROOT}/scripts/run.sh" promptran

View source on GitHub

More from Ajw2003/AjsClaudeCodeTools

All 3 plugins →

STATUS: shell/v0.1. Classifies each prompt's complexity and nudges Claude to delegate to a model-pinned subagent - haiku for simple doc/comms work, sonnet for recon and code implementation, opus for planning, architecture, and management. Cannot switch the live session's own model.

1 command · 3 agents · hooks

1

STATUS: shell/v0.1. Notices an under-specified prompt and walks the user through goal, constraints, success criteria and gating before Claude runs with an inferred reading of it.

1 command · hooks

1

Maintain house-rules?

This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.

claude & codex plugins3 plugins