CleanExpo/.claude/skills/next
.claude/skills/next/SKILL.md
# Next Skill **Name:** next **Triggers:** `/next` (manual invocation only — not natural language) **Version:** 1.0.0 --- ## Purpose Task transition. Snapshots the current task state to disk, then guides the user to either `/compact` (if the next task is related) or `/clear` (if unrelated). Single biggest token saver in the optimization pack — the audit found that 98.5% of tokens in long sessions are conversation re-reading, and `/clear` between unrelated tasks eliminates that cost entirely.
$ install --global
skillsauthnpx skillsauth add CleanExpo/CCW-CRM .claude/skills/nextInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
Security Scan Results
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
3
Scanners Passed
9
Scanners in report
Clean
TrivyContainer and dependency vulnerability scanner
95%
Clean
SemgrepStatic code analysis for vulnerabilities
95%
Clean
mcp-scan (Snyk)Model Context Protocol security validation
95%
Skipped
Snyk (dep)Open source security scanning
50%
Skipped
Socket.devSupply chain security analysis
50%
Skipped
VirusTotalMulti-engine malware detection
50%
Skipped
CrowdStrikeAdvanced threat intelligence
50%
Skipped
OSV-ScannerOpen Source Vulnerability database check
50%
Skipped
OWASP Dep-Check
50%
Last scanned: Apr 16, 2026, 12:50 PM202.3s1 file scanned
SKILL.md
Next Skill
Name: next
Triggers: /next (manual invocation only — not natural language)
Version: 1.0.0
Purpose
Task transition. Snapshots the current task state to disk, then guides the user to either /compact (if the next task is related) or /clear (if unrelated). Single biggest token saver in the optimization pack — the audit found that 98.5% of tokens in long sessions are conversation re-reading, and /clear between unrelated tasks eliminates that cost entirely.
This skill is manual invocation only. It is not triggered by natural language because the decision to compact or clear is intentional and CEO-driven.
Commands
/next
Runs three steps:
Step 1 — Snapshot. Writes .claude/snapshots/last-task.md containing:
- What completed (checklist of done items)
- Files modified (full paths)
- Decisions made (architectural choices, naming conventions, gotchas discovered)
- Next steps (what the user said they want to do next, even if vague)
The snapshot is plain markdown. /resume reads it later to recover state without re-reading the conversation.
Step 2 — Ask. Asks the user one question via AskUserQuestion:
Is your next task RELATED to what we just did, or UNRELATED?
- RELATED — same module, same file, follow-up work → run
/compactand preserve file paths + decisions + test results- UNRELATED — different module, different concern → run
/clearand start fresh
Step 3 — Confirm. After the user picks, the skill prints one line:
Ready. Context [compacted/cleared]. Run /resume to restore the snapshot if needed.
Snapshot Format
.claude/snapshots/last-task.md:
# Last Task Snapshot
**Date:** YYYY-MM-DD HH:MM
**Branch:** claude/<worktree>
**Status:** [in-progress | complete | blocked]
## Completed
- [item 1]
- [item 2]
## Files Modified
- apps/backend/src/api/routes/X.py
- apps/web/components/Y.tsx
## Decisions
- [architectural choice + why]
- [pattern used + reference file]
## Next Steps
- [what the user asked for next]
- [open questions]
## Verification State
- Tests: [passing | failing | not run]
- Lint: [clean | dirty | not run]
- Type-check: [clean | dirty | not run]
Process Rules
- Snapshot before asking. Write the file first, THEN ask the user. If the user closes the session before answering, the snapshot is still on disk.
- Never auto-clear. The user must explicitly choose.
/clearis destructive to conversation context. - Never auto-compact. Same reason. The user controls when context is reduced.
- Disable model invocation. This skill must never fire on natural language — only the explicit
/nextslash command. - Snapshot is plain markdown. No frontmatter, no YAML, no JSON.
/resumeparses it as plain text.
When NOT to use this skill
- You're mid-task and not done — finish first, then
/next - You don't plan to start a new task — just leave the session open
- You want to fully wipe state including the snapshot —
/cleardirectly
Related
.claude/skills/resume/SKILL.md— reads the snapshot this skill writes.claude/skills/ctx/SKILL.md— diagnoses whether/compactor/clearis needed.claude/snapshots/— directory that holds snapshots and pre-compact diffs (auto-created by hooks + this skill)
Related Skills
CleanExpo/youtube-uploader
content-media
VerifiedTrustedCommunity
Autonomously uploads CCW HeyGen demo videos to YouTube as Unlisted, collects video IDs, and patches DemoVideoBanner.tsx + video-registry.json. One-time OAuth setup required. Handles resume, retries, and ID propagation.
1SKILL.mdUpdated Apr 16, 2026
CleanExpo/unfreeze
data-ai
VerifiedTrustedCommunity
Clear the freeze boundary set by /freeze, allowing edits to all directories again. Use when you want to widen edit scope without ending the session. Use when asked to "unfreeze", "unlock edits", "remove freeze", or "allow all edits". (gstack)
1SKILL.mdUpdated Apr 16, 2026
CleanExpo/.claude/skills/spec-interview
tools
VerifiedTrustedCommunity
# Spec Interview Skill **Name:** spec-interview **Triggers:** `/spec-interview`, when requirements unclear **Version:** 1.0.0 --- ## Purpose Interviews user to gather complete requirements before planning. --- ## Interview Questions When requirements are unclear, ask: ### 1. Feature Clarity **Question:** "What is this feature supposed to do?" **Why:** Need clear objective ### 2. User Impact **Question:** "Who will use this and why?" **Why:** Understand user needs ### 3. Success Criteri
1SKILL.mdUpdated Apr 16, 2026
CleanExpo/ship
development
VerifiedTrustedCommunity
Ship workflow: detect + merge base branch, run tests, review diff, bump VERSION, update CHANGELOG, commit, push, create PR. Use when asked to "ship", "deploy", "push to main", "create a PR", "merge and push", or "get it deployed". Proactively invoke this skill (do NOT push/PR directly) when the user says code is ready, asks about deploying, wants to push code up, or asks to create a PR. (gstack)
1SKILL.mdUpdated Apr 16, 2026