How do I install the Wrap Up skill?

Run npx skillmds@latest add skogai/wrap-up in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

What does the Wrap Up skill do?

end-of-session checklist — ship, remember, review, journal. invoke at session close. It is listed under Coding & Dev Tools on SkillMD.

Is Wrap Up safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Wrap Up?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Wrap Up free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Wrap Up?

skogai (@skogai) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Wrap Up by skogai — Agent Skill

Name: Wrap Up
Author: skogai

File contents

Run four phases in order. Auto-apply all findings without asking; present a consolidated report at the end.

phase 1: ship

task sync (if gptodo is on PATH):

gptodo fetch --all — refresh external issue states
gptodo list — find stale in-progress tasks; mark completed ones with gptodo edit <id> --state done

commit:

git status on every repo touched this session
Commit uncommitted changes with a descriptive message, push to remote

worktree cleanup (if wt is on PATH):

wt list — merge and remove completed worktrees

file placement check:

For any files created or saved this session: verify naming follows kebab-case convention
Auto-fix naming violations (rename the file)
If a .md file isn't a {CLAUDE,SKOGAI,AGENTS}.md, it should live in docs/ unless explicitly placed elsewhere — auto-move misplaced files

phase 2: remember

Review what was learned. Place each piece in the right location:

auto-memory (.planning/memory/) — debugging insights, patterns, project quirks
CLAUDE.md — permanent conventions or commands that should guide all future sessions
.claude/rules/ — topic-specific rules scoped to file types (use paths: frontmatter)
@import reference — when CLAUDE.md would benefit from referencing another file instead of duplicating content
gptodo task — future work that needs research or has dependencies

phase 3: review & apply

Analyze the conversation for self-improvement opportunities. If the session was short or routine, say "nothing to improve" and proceed.

categories:

skill gap — things that were wrong or needed multiple attempts
friction — repeated manual steps that should be automatic
knowledge — facts about projects or preferences that should have been known
automation — repetitive patterns that could become skills, hooks, or scripts

Auto-apply all actionable findings. Present a summary:

findings (applied):
1. <category>: <what happened>
   → [location] <what was changed>

no action needed:
1. <what was found> — already documented

phase 4: journal

Review the full conversation for material worth preserving: interesting solutions, milestones, architectural decisions, debugging stories.

If journal-worthy: save to personal/journal/YYYY-MM-DD/<description>.md (append-only).

Otherwise: "nothing worth journaling from this session."

---name: wrap-updescription: end-of-session checklist — ship, remember, review, journal. invoke at session close.---Run four phases in order. Auto-apply all findings without asking; present a consolidated report at the end.## phase 1: ship**task sync** (if `gptodo` is on PATH):1. `gptodo fetch --all` — refresh external issue states2. `gptodo list` — find stale in-progress tasks; mark completed ones with `gptodo edit <id> --state done`**commit:**1. `git status` on every repo touched this session2. Commit uncommitted changes with a descriptive message, push to remote**worktree cleanup** (if `wt` is on PATH):1. `wt list` — merge and remove completed worktrees**file placement check:**1. For any files created or saved this session: verify naming follows kebab-case convention2. Auto-fix naming violations (rename the file)3. If a `.md` file isn't a `{CLAUDE,SKOGAI,AGENTS}.md`, it should live in `docs/` unless explicitly placed elsewhere — auto-move misplaced files## phase 2: rememberReview what was learned. Place each piece in the right location:- **auto-memory** (`.planning/memory/`) — debugging insights, patterns, project quirks- **CLAUDE.md** — permanent conventions or commands that should guide all future sessions- **.claude/rules/** — topic-specific rules scoped to file types (use `paths:` frontmatter)- **`@import` reference** — when CLAUDE.md would benefit from referencing another file instead of duplicating content- **gptodo task** — future work that needs research or has dependencies## phase 3: review & applyAnalyze the conversation for self-improvement opportunities. If the session was short or routine, say "nothing to improve" and proceed.**categories:**- **skill gap** — things that were wrong or needed multiple attempts- **friction** — repeated manual steps that should be automatic- **knowledge** — facts about projects or preferences that should have been known- **automation** — repetitive patterns that could become skills, hooks, or scriptsAuto-apply all actionable findings. Present a summary:```findings (applied):1. <category>: <what happened> → [location] <what was changed>no action needed:1. <what was found> — already documented```## phase 4: journalReview the full conversation for material worth preserving: interesting solutions, milestones, architectural decisions, debugging stories.If journal-worthy: save to `personal/journal/YYYY-MM-DD/<description>.md` (append-only).Otherwise: "nothing worth journaling from this session."

Frequently asked questions

Wrap Up