kenmark-skills 1.0.0

package/README.md

@@ -0,0 +1,115 @@
+# kenmark-skills
+
+Agent skills from [Kenmark ITan Solutions](https://github.com/TanoojMehra/kenmark-skills) — reusable workflows for **Cursor**, **Codex**, **Claude Code**, and other harnesses that read `SKILL.md` files.
+
+Created by **Tanooj Mehra** and **Adwait Date**.
+
+## What's included
+
+| Skill | Purpose |
+| --- | --- |
+| [`init-brain`](skills/user-skills/init-brain/SKILL.md) | Bootstrap `brain/` and sync project standards to `CLAUDE.md`, `AGENTS.md`, `.cursorrules`, and `.cursor/rules/` |
+| [`skill-router`](skills/user-skills/skill-router/SKILL.md) | Pick the best installed user skill from intent and keywords |
+| [`commit-push`](skills/user-skills/commit-push/SKILL.md) | Group changes by feature, write conventional commits, push (no co-author trailers) |
+| [`issues-setup`](skills/user-skills/issues-setup/SKILL.md) | Create `brain/issues/` layout and `INDEX.md` |
+| [`issues-list`](skills/user-skills/issues-list/SKILL.md) | Dashboard of open issues by area and priority |
+| [`issues-check`](skills/user-skills/issues-check/SKILL.md) | Move resolved issues to `completed/` and refresh the index |
+| [`issues-scan`](skills/user-skills/issues-scan/SKILL.md) | Discover new issues in the codebase and file them under `brain/issues/` |
+| [`issues-maintenance`](skills/user-skills/issues-maintenance/SKILL.md) | Audit issue tracker health (duplicates, stale index, missing files) |
+
+These skills assume a **`brain/`** knowledge base (standards, changelog, optional issue tracker). Use **`init-brain`** first on a new repo, then the issues and git workflow skills as needed.
+
+## Requirements
+
+- **Node.js** 18+ (for npm install/publish only)
+- An agent environment that discovers skills from disk (Cursor, Codex CLI, Claude Code, etc.)
+
+## Install
+
+### Option A — Clone into your skills directory
+
+**Cursor (user-level):**
+
+```bash
+git clone https://github.com/TanoojMehra/kenmark-skills.git
+cp -R kenmark-skills/skills/user-skills/* ~/.cursor/skills/
+```
+
+**Codex / shared agent skills (`~/.agents/skills`):**
+
+```bash
+cp -R kenmark-skills/skills/user-skills/* ~/.agents/skills/
+```
+
+**Project-scoped (team repo):**
+
+```bash
+mkdir -p .cursor/skills
+cp -R /path/to/kenmark-skills/skills/user-skills/* .cursor/skills/
+```
+
+Commit project-scoped skills only if the whole team should use the same versions.
+
+### Option B — npm package
+
+```bash
+npm install kenmark-skills
+```
+
+Published files live under `node_modules/kenmark-skills/skills/user-skills/`. Copy that folder into `~/.cursor/skills/`, `~/.agents/skills/`, or your project's skills path (same as Option A).
+
+Verify what will ship before publishing:
+
+```bash
+npm run pack:check
+```
+
+Publish (maintainers):
+
+```bash
+npm run publish:public
+```
+
+## Using a skill
+
+1. Ensure the skill folder (e.g. `init-brain/`) containing `SKILL.md` is on the agent's skill search path.
+2. In chat, refer to the skill by name or trigger phrase, for example:
+   - "Run **init-brain** on this repo"
+   - "Use **commit-push** to ship these changes"
+   - "**issues-list**" or "list open issues"
+3. The agent should **read and follow** the full `SKILL.md` for that skill — not improvise from the description alone.
+
+### skill-router note
+
+`skill-router` builds a local registry at `skills/user-skills/skill-router/registry.json` when run from a checkout of this repo. That file is **gitignored** and **not** published to npm; it is per-machine generated state.
+
+## Repository layout
+
+```text
+kenmark-skills/
+├── README.md
+├── package.json
+├── brain/
+│   └── CHANGELOG.md          # Versioned notes for this package
+└── skills/
+    └── user-skills/          # Skills shipped in the npm package
+        ├── init-brain/
+        ├── skill-router/
+        ├── commit-push/
+        └── issues-*/
+```
+
+## Contributing
+
+1. Edit skills under `skills/user-skills/<skill-name>/SKILL.md`.
+2. Append an entry to [`brain/CHANGELOG.md`](brain/CHANGELOG.md) with a dated version slug.
+3. Open a PR; describe which skills changed and any new trigger phrases.
+
+## License
+
+MIT — Kenmark ITan Solutions (see `package.json`).
+
+## Links
+
+- **Repository:** https://github.com/TanoojMehra/kenmark-skills
+- **Issues:** https://github.com/TanoojMehra/kenmark-skills/issues

package/brain/CHANGELOG.md

@@ -0,0 +1,48 @@
+# CHANGELOG
+
+## v2026.06.02-1912-init-brain-user-choice
+- init-brain now asks which agent config files to sync instead of creating all targets by default.
+- Brain scaffold still always runs; only user-selected files (CLAUDE.md, AGENTS.md, Cursor, GEMINI.md) are written.
+
+## v2026.06.02-readme
+- Added root `README.md` with skill inventory, install paths (Cursor, `~/.agents/skills`, npm), usage notes, and repo layout.
+
+## v2026.06.02-package-metadata
+- Updated `package.json` description, authors, and GitHub URLs for Kenmark ITan Solutions (`TanoojMehra/kenmark-skills`).
+
+## v2026.06.02-1905-init-brain-cross-agent
+- Expanded `skills/user-skills/init-brain/SKILL.md` to bootstrap `brain/` and sync standards to `CLAUDE.md`, `AGENTS.md`, `.cursorrules`, and `.cursor/rules/project-standards.mdc`.
+- Documented `brain/rules/standards.md` as source of truth with `init-brain` merge markers for idempotent updates across Claude Code, Codex, and Cursor.
+
+## v2026.06.02-1810-init-skills
+- Initialized project scaffold with local `skills/`, `brain/`, and `temp/` directories.
+- Copied user-authored skills from `~/.claude/skills` and `~/.codex/skills` into workspace-local skill packs.
+
+## v2026.06.02-1814-skill-filter
+- Filtered copied skills to requested set only: `issues-*`, `commit-push`, and `skill-router`.
+- Re-copied using symlink target resolution so requested skills are fully included.
+
+## v2026.06.02-1819-skill-unify
+- Merged split skill packs into a single folder: `skills/user-skills/`.
+- Deduped by skill name and preserved source provenance in `skills/user-skills-manifest.md`.
+- Removed `skills/claude-user-skills/` and `skills/codex-user-skills/` after merge.
+
+## v2026.06.02-1820-brain-rules-skill
+- Added `skills/user-skills/brain-rules/SKILL.md` with all provided Cursor rules embedded in one reusable skill.
+- Preserved all rule sections and bullet points from the supplied rule set.
+
+## v2026.06.02-1826-npm-package-config
+- Added root `package.json` for npm publishing with public access config and Node engine requirement.
+- Whitelisted publish files to include unified user skills and changelog artifacts only.
+- Added helper scripts: `pack:check` and `publish:public`.
+
+## v2026.06.02-1844-init-brain-rename
+- Renamed skill folder from `skills/user-skills/brain-rules/` to `skills/user-skills/init-brain/`.
+- Updated skill title to `Init Brain Skill`.
+- Updated skill inventory files to include `init-brain` and refresh counts.
+
+## v2026.06.02-1851-skill-router-registry-runtime
+- Updated `skills/user-skills/skill-router/SKILL.md` to generate `skills/user-skills/skill-router/registry.json` at runtime on each invocation.
+- Added repo and publish ignores so the generated registry file is not committed or shipped (`.gitignore`, `.npmignore`).
+- Removed tracked `skills/user-skills/skill-router/registry.json`; it is now per-user generated state.
+

package/package.json

@@ -0,0 +1,46 @@
+{
+  "name": "kenmark-skills",
+  "version": "1.0.0",
+  "description": "Skills by Kenmark ITan Solutions — Cursor/Codex agent skills from our development workflows. Created by Tanooj Mehra & Adwait Date.",
+  "license": "MIT",
+  "private": false,
+  "type": "commonjs",
+  "files": [
+    "skills/user-skills/**/*",
+    "skills/user-skills-manifest.md",
+    "skills/COPIED_SKILLS.md",
+    "brain/CHANGELOG.md"
+  ],
+  "scripts": {
+    "pack:check": "npm pack --dry-run",
+    "publish:public": "npm publish --access public"
+  },
+  "keywords": [
+    "kenmark",
+    "cursor",
+    "codex",
+    "skills",
+    "ai-agent",
+    "workflow",
+    "rules"
+  ],
+  "author": "Kenmark ITan Solutions",
+  "contributors": [
+    "Tanooj Mehra",
+    "Adwait Date"
+  ],
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/TanoojMehra/kenmark-skills.git"
+  },
+  "bugs": {
+    "url": "https://github.com/TanoojMehra/kenmark-skills/issues"
+  },
+  "homepage": "https://github.com/TanoojMehra/kenmark-skills#readme",
+  "publishConfig": {
+    "access": "public"
+  },
+  "engines": {
+    "node": ">=18"
+  }
+}

package/skills/user-skills/commit-push/SKILL.md

@@ -0,0 +1,225 @@
+---
+name: commit-push
+description: |
+  Systematically group working-tree changes by feature or area, create meaningful
+  conventional commits, and push to the remote. Never adds Co-authored-by or other
+  attribution trailers. Use when asked to "commit and push", "commit by feature",
+  "split commits", "ship my changes", "commit logically", or "push commits".
+allowed-tools:
+  - Bash
+  - Read
+  - Grep
+  - Glob
+  - AskUserQuestion
+triggers:
+  - commit and push
+  - commit by feature
+  - split commits
+  - ship my changes
+  - commit logically
+  - push commits
+  - systematic commit
+---
+
+# Commit & Push — Feature-Grouped, Meaningful Commits
+
+## Purpose
+
+Turn a messy working tree into one or more **focused commits** grouped by feature
+or area, write messages that explain **why**, push to the tracking remote — and
+**never** add co-author or tool attribution.
+
+Only run when the user explicitly asks to commit and/or push. Do not commit
+opportunistically after unrelated edits.
+
+---
+
+## Hard rules (non-negotiable)
+
+1. **No co-author attribution**
+   - Never add `Co-authored-by:`, `Made-with:`, `Signed-off-by:` for an AI/tool, or
+     similar trailers to commit messages.
+   - Never pass `--trailer` with co-author metadata.
+   - If a hook rewrites the message to add co-author lines, **abort the commit**,
+     fix or bypass the hook only if the user explicitly requests it, and retry
+     with a clean message.
+   - After each commit, verify with `git log -1 --format=%B` that no attribution
+     trailer was injected.
+
+2. **No git config changes** — do not run `git config` (local or global).
+
+3. **No destructive git** — no `push --force`, `reset --hard`, or skipping hooks
+   (`--no-verify`) unless the user explicitly asks.
+
+4. **No push to main/master by default** — if on the default branch, create a
+   feature branch first (derive name from the dominant change), then commit and
+   push with `-u origin HEAD`.
+
+5. **No secrets in commits** — never stage `.env`, credentials, keys, or token
+   files. Warn if the user asked to commit them.
+
+---
+
+## Step 1 — Gather context (parallel)
+
+Run these in parallel:
+
+```bash
+git status --short
+git diff --stat
+git diff --cached --stat
+git branch -vv
+git log --oneline -12
+git rev-parse --abbrev-ref HEAD
+git rev-parse --abbrev-ref origin/HEAD 2>/dev/null || true
+```
+
+Resolve default branch: strip `origin/` from `origin/HEAD`, or use `main`.
+
+If the working tree is clean, report nothing to commit and stop.
+
+---
+
+## Step 2 — Group changes by feature / area
+
+Partition **staged + unstaged + untracked** files into logical commit groups.
+Prefer **file-level** grouping (do not use `git add -p`).
+
+### Monorepo grouping (cercli-integrations)
+
+| Group key | Paths | Typical type |
+|-----------|-------|--------------|
+| `web` | `apps/web/**` | feat/fix/test |
+| `worker` | `apps/worker/**` | feat/fix/test |
+| `packages/<name>` | `packages/<name>/**` | feat/fix/test |
+| `brain` | `brain/**` | docs/chore |
+| `db` | `packages/db/**`, migrations | feat/fix |
+| `root` | root configs, `ecosystem.config.cjs`, CI | chore/ci |
+| `agent` | `packages/agent/**` | feat/fix |
+| `dsl` | `packages/dsl/**` | feat/fix |
+
+Rules:
+
+- **Tests** go with the code they cover (same commit group).
+- **Renames/moves** stay with the feature they belong to.
+- **Unrelated concerns** → separate commits (sweet spot: 2–4 commits; avoid
+  10+ micro-commits).
+- If one file mixes unrelated concerns, commit it with the **dominant** group and
+  note the remainder for a follow-up — do not hunk-split.
+
+Output a short plan before committing:
+
+```
+Commit plan:
+1. feat(web): … — apps/web/src/… (N files)
+2. test(worker): … — apps/worker/src/__tests__/… (M files)
+3. docs(brain): … — brain/CHANGELOG.md (K files)
+```
+
+Use **AskUserQuestion** (or ask in chat) only when grouping is ambiguous and
+stakes are high (e.g., half-finished feature vs ready-to-ship).
+
+---
+
+## Step 3 — Message convention
+
+Priority:
+
+1. Project docs (`CLAUDE.md`, `AGENTS.md`) if loaded and they specify format.
+2. Recent `git log` on this branch — match dominant pattern.
+3. Default: **Conventional Commits** — `type(scope): imperative subject`
+
+Types: `feat`, `fix`, `docs`, `refactor`, `test`, `chore`, `perf`, `ci`, `build`
+
+- **Subject**: imperative, ≤72 chars, explains **why** / user impact, not a file list.
+- **Body**: optional blank line + 1–3 sentences for non-obvious motivation,
+  trade-offs, or issue IDs. Omit for trivial single-purpose changes.
+
+Examples from this repo:
+
+```
+feat(web): wire connections UI to credentials API
+fix(web): validate workflow DSL against credential IDs
+test(web): align Prisma fakes with Credential model
+docs(brain): credential migration audit register
+```
+
+Bad:
+
+```
+update files
+fix stuff
+WIP
+feat: changes (co-authored by …)   ← NEVER
+```
+
+---
+
+## Step 4 — Commit each group (sequential)
+
+For each group:
+
+```bash
+git add path/to/file1 path/to/file2 ...
+git commit -m "$(cat <<'EOF'
+type(scope): subject focused on why
+
+Optional body when the change is not self-evident.
+EOF
+)"
+git log -1 --format=%B   # verify NO Co-authored-by / Made-with trailers
+```
+
+If pre-commit hook **modifies** files: fix issues, `git add` the hook changes,
+create a **new** commit (do not `--amend` unless user rules allow amend).
+
+If commit **fails**: read hook output, fix, new commit — never amend a failed commit.
+
+After all groups:
+
+```bash
+git status
+```
+
+Working tree should be clean (or only intentionally excluded files remain).
+
+---
+
+## Step 5 — Push
+
+Only push when this skill was invoked (user asked to commit **and** push) or
+explicitly said "push".
+
+```bash
+# Ensure upstream exists
+git rev-parse --abbrev-ref @{upstream} 2>/dev/null || git push -u origin HEAD
+git push
+```
+
+If push is rejected (non-fast-forward):
+
+- Report the error; suggest `git pull --rebase` only if the user asks.
+- **Never** force-push to `main`/`master`.
+
+Report: branch name, commit SHAs + subjects, remote URL if useful, push result.
+
+---
+
+## Step 6 — Post-commit hygiene (cercli-integrations)
+
+If code changed (not docs-only), ensure `brain/CHANGELOG.md` reflects the work
+when project rules require it. Either include `brain/` in the relevant commit or
+add a small follow-up `docs(brain): …` commit in the same push batch.
+
+---
+
+## Quick reference — do / don't
+
+| Do | Don't |
+|----|--------|
+| Group by feature/package | One giant "misc" commit |
+| `git add` explicit paths | `git add -A` blindly |
+| Verify message has no trailers | Co-authored-by anywhere |
+| Feature branch off main | Commit directly on main |
+| `git push -u origin HEAD` first time | Force push |
+| HEREDOC for messages | `-m "line1\nline2"` escaping mistakes |