Skills for Real Engineers. Straight from my .agents directory.
review-cockpit
Agent skill fromhuytieu/COG-second-brain
Produce and continuously maintain ONE living review document for a multi-item session — a cockpit header (Progress checklist, Working folder, Context) plus per-item review cards that you approve or request changes on directly in the doc or side panel. Use whenever a session has multiple deliverables you need to review/approve (meeting-processing + planning, multi-ticket work, briefs with several drafts, any "do X, then plan/draft Y and Z"). The doc is the interaction surface: you edit it (or reply in chat) to approve/change; the agent keeps it live as work progresses.
How to install review-cockpit
npx skills add huytieu/COG-second-brain# or add only this skill to Claude Code: copy the folder
skills/review-cockpit/
# into
~/.claude/skills/review-cockpit/ (all projects)
.claude/skills/review-cockpit/ (one project)
1# review-cockpit23## Purpose45A single, living **review document** that doubles as the control surface for a session. Instead of scattering a meeting note here, a ticket there, two drafts in chat, everything lands in **one file** you can open in the Obsidian side panel (or Claude side panel) and drive: see progress at a glance, review each item in place, approve or request changes inline, and watch the doc update as the agent works. Mirrors a "co-work" cockpit (Progress / Working folder / Context).67## When to use89- Any session with **≥2 things you need to review or approve** (the default for "process X then plan/draft Y, Z").10- Meeting-processing + pod planning, multi-ticket runs, briefs with several drafts, spec + plan + drafts.11- NOT for a single trivial deliverable (one file, no approval loop) — a plain file is fine there.1213Pairs with `closed-loop` (verification), `harvest` (learnings), and the V-model checkpoints. This skill governs **how the deliverable is presented and driven**, not the verification pipeline.1415## The one rule1617**One doc per session.** It opens with a cockpit, then one review card per item. Every artifact the session produces is linked from the **Working folder** table (fan-out to sub-files mid-run is fine; the cockpit is the single front door). Never hand the user "see files A, B, C" — hand them the cockpit.1819## Structure20…
Excerpt shown under the repository's MIT license.
More skills in COG-second-brain
Similar skill repositories
An agentic skills framework & software development methodology that works.
A single CLAUDE.md file to improve Claude Code behavior, derived from Andrej Karpathy's observations on LLM coding pitfalls.
Public repository for Agent Skills
Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.
Persistent Context Across Sessions for Every Agent – Captures everything your agent does during sessions, compresses it with AI, and injects relevant context back into future sessions. Works with Claude Code, OpenClaw, Codex, Gemini, Hermes, Copilot, OpenCode + More
