docs: fix README skill and hook counts and group tables by purpose

[EdbertChan]

Summary

The project's front page lists every skill and hook, with counts at the top. The counts were wrong, some hooks were missing, and the lists were long flat tables.

This change fixes the counts, adds the missing hooks, and groups skills and hooks by what they are for.

Review Claim

The front page and inventory page now match the skills and hooks in the repository, grouped by purpose.

Review Lane

docs

Review Unit

docs

Safety Invariant

Documentation only: no hook, skill, installer, or test file changes, so nothing that runs on install or in a session can change behavior.

Slice Rationale

The front page and the inventory page describe the same inventory, so they move together. The inventory page rows ride along per the review-unit table.

Non-goals

• No change to any hook or skill.
• No change to which skills install where.

Test Plan

Test Plan

• find engine/skills corpus/skills product/skills -mindepth 2 -maxdepth 2 -name SKILL.md | wc -l → engine 7, corpus 32, product 18 (57), matching the badge and section headers
• ls engine/hooks | grep -v '^_' | grep -v '\.' | wc -l → 53, matching the badge, the hooks paragraph, and the section counts (7+6+5+5+5+11+2+12)
• every hook folder appears in the README hook tables (comm of folder list vs table rows is empty both ways)
• every hook and skill name in the tables exists under engine/hooks/ or */skills/
• grep -c '^||' docs/ecosystem.md → 0

Revert Plan

Revert Plan

• Safe to revert? Yes
• Revert command: git revert <sha>
• Post-revert steps: None
• Data migration? No

🤖 Generated with Claude Code

---

Note

Low Risk
Documentation-only changes to README and ecosystem inventory; no install, hook, or skill runtime behavior is modified.

Overview
Updates README marketing badges and restructures skills/hooks documentation so counts match the current tree (57 skills, 53 hooks) and readers can scan by purpose instead of one flat table.

The Skills section is split into engine (7), corpus (32), and product (18) with sub-headings (mining, PR workflow, evidence, etc.); several skills that were missing or buried in a single list (make-pr, phrase-judge, plan-first, prove-it-ship-gate, and others) are now named explicitly. The Hooks section grows from a short sample to categorized tables (evidence, session hygiene, routing, Git/PR, code quality, opt-in reflect hooks, unverified-tag pair, plus a catch-all bucket), with a pointer to docs/ecosystem.md for the canonical full list and tighter one-line descriptions (e.g. reflect hooks gated on CATSTACK_REFLECT_ENFORCEMENT).

docs/ecosystem.md adds engine inventory rows for hooks that were live but not listed (unverified-tag-ledger, gh-write-verification, no-comments, split-scope / narrow-the-scope injectors, and several judge/evidence hooks).

^{Reviewed by Cursor Bugbot for commit b19d512. Bugbot is set up for automated code reviews on this repo. Configure here.}

[cursor]

Bugbot couldn't run - usage limit reached

Bugbot is counted against Cursor usage for this user or team, and this run hit a usage or spend limit.

A user or team admin can review and increase usage limits in the Cursor dashboard.

(requestId: serverGenReqId_73d6393e-056c-4c58-b124-160d9da75484)

[mergify]

Queued — the merge queue status continues in this comment ↓.

[EdbertChan]

@Mergifyio queue

[mergify]

Merge Queue Status

• ✅ Entered queue — 2026-09-28 16:29 UTC · Rule: admin-bypass · triggered by @EdbertChan with the @mergifyio queue command
• ✅ Checks passed · on draft merge queue: checking #1214 on main (ff035b8) #1215
• ✅ Merged — 2026-09-28 17:14 UTC · at 0aa9224735f200fae97719cdff337f73a1f51726 · squash

This pull request spent 44 minutes 27 seconds in the queue, including 43 minutes 59 seconds running CI.

Required conditions to merge

• check-success = lint
• check-success = test
• check-success = validate