From 9858611afcdf848b6e483ba78f89c576aa5c8743 Mon Sep 17 00:00:00 2001 From: Craig Jennings Date: Thu, 7 May 2026 09:36:14 -0500 Subject: chore(skills): remove humanizer (superseded by voice) MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit I deleted humanizer/SKILL.md now that all three callers (commits.md, respond-to-cj-comments.md, start-work.md) invoke /voice instead. The 25 humanizer patterns live on as patterns 1-25 in voice/SKILL.md. Same source (Wikipedia's Signs of AI writing), same prose, same examples — just renumbered alongside the universal good-writing additions and the personal-only patterns. I also updated .ai/notes.org and .ai/workflows/wrap-it-up.org to reference /voice personal instead of the old humanizer + manual-passes flow. The wrap-it-up change landed upstream in claude-templates first so it survives the next startup rsync. todo.org gets the matching update: the voice TODO is marked DONE with a "Built and shipped" timestamp, the publish-mode terminology is renamed to personal-mode throughout, the V1 scope checklist is ticked, the open questions are resolved with the answers we landed on during implementation, and the migration section records the delete-not-alias decision. --- .ai/notes.org | 2 +- .ai/workflows/wrap-it-up.org | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) (limited to '.ai') diff --git a/.ai/notes.org b/.ai/notes.org index 735e256..0c75bb1 100644 --- a/.ai/notes.org +++ b/.ai/notes.org @@ -31,7 +31,7 @@ See [[file:../README.org][README.org]] for the full layout, install modes, and l ** Layout (high level) - =claude-rules/= — generic rules (=commits.md=, =testing.md=, =verification.md=, =subagents.md=) symlinked into =~/.claude/rules/= and applied to every Claude Code session on the machine. -- Top-level skill directories (=add-tests/=, =debug/=, =five-whys/=, =frontend-design/=, =humanizer/=, =pairwise-tests/=, =playwright-js/=, =playwright-py/=, =root-cause-trace/=) — each a Claude Code skill, symlinked into =~/.claude/skills/= by =make install=. +- Top-level skill directories (=add-tests/=, =debug/=, =five-whys/=, =frontend-design/=, =pairwise-tests/=, =playwright-js/=, =playwright-py/=, =root-cause-trace/=, =voice/=) — each a Claude Code skill, symlinked into =~/.claude/skills/= by =make install=. - =languages/= — per-language bundles (rules + hooks + settings) copied into target projects via =make install-lang LANG= PROJECT==. Both =LANG= and =PROJECT= are optional — fzf picks them interactively when omitted. Bundles currently shipping: =elisp=, =python=. - =.claude/= — repo-local Claude Code config: =settings.json= and =commands/=. - =hooks/=, =scripts/= — install helpers and PostToolUse validators that ride along with bundles. diff --git a/.ai/workflows/wrap-it-up.org b/.ai/workflows/wrap-it-up.org index 37d2522..5c5dc89 100644 --- a/.ai/workflows/wrap-it-up.org +++ b/.ai/workflows/wrap-it-up.org @@ -119,7 +119,7 @@ Commit message rules (also see protocols.org "Git Commit Requirements"): *Wrap-up commits skip the inline-approval gate.* The =commits.md= rule that requires writing the message to =/tmp/commit-.md=, printing inline, and waiting for an approve / request-changes / open-in-editor response does *not* apply to wrap-up commits. The wrap-up flow is meant to be quick — Craig has already authorized the wrap by triggering the workflow ("wrap it up"), and stopping again to approve a commit message disrupts the cadence. -Still apply the prose-quality passes silently before committing — humanizer + jargon-rewrite + semicolon-swap + contractions + sentence-split — so the message reads cleanly. Just don't print and ask. Commit directly with the cleaned message. +Still apply =/voice personal= silently before committing so the message reads cleanly. Just don't print and ask. Commit directly with the cleaned message. If a wrap-up commit needs Craig's eyes for a content reason (sensitive change, unusual scope, something he flagged earlier), surface it explicitly. Otherwise commit and move on. -- cgit v1.2.3