# Closing

> Wrap up the session before closing — persist memory, refresh CLAUDE.md, flag loose ends

- **Type:** Skill
- **Install:** `agentstack add skill-wecko-ai-claude-skills-closing`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [Wecko-ai](https://agentstack.voostack.com/s/wecko-ai)
- **Installs:** 0
- **Category:** [Agent Skills](https://agentstack.voostack.com/c/agent-skills)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [Wecko-ai](https://github.com/Wecko-ai)
- **Source:** https://github.com/Wecko-ai/claude-skills/tree/main/plugins/closing

## Install

```sh
agentstack add skill-wecko-ai-claude-skills-closing
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

The user is about to close this session. Do a full wrap-up so nothing is lost and the next session starts clean. Work through ALL of these, then give a short final report:

## 1. Memory (auto-memory directory)
- Review what happened this session: decisions, gotchas discovered, fixes, project state changes, user feedback/corrections.
- For each durable, non-obvious fact: update the existing memory file if one covers it, else create a new one (correct frontmatter, `[[links]]`).
- Delete or rewrite memories this session proved WRONG or outdated (e.g. "X is broken" after X was fixed).
- Sync `MEMORY.md` index: one line per memory, hooks accurate, no orphan or stale entries.

## 2. CLAUDE.md freshness
- Diff what changed this session (architecture, commands, workflows, deploys, env) against project CLAUDE.md claims.
- If CLAUDE.md is stale: small focused update. If the repo requires PRs for changes, do branch + PR; if docs-only commits are fine, commit directly per project rules.

## 3. Loose ends — report, don't fix
- `git status`: uncommitted/untracked files — say which are from this session vs foreign (other tools/sessions). Never absorb foreign changes.
- Unmerged branches / open PRs created this session and their CI state.
- Background tasks/monitors still running (kill or hand off).
- Anything half-done or promised but not delivered — list explicitly with what remains.
- Known pending chores discovered this session (note them in memory if not already).

## 4. Final report
Short summary: what was persisted (memory files touched), CLAUDE.md updated or not, loose ends list. No fluff.

If the user passed notes, treat them as priorities for the wrap-up: $ARGUMENTS

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [Wecko-ai](https://github.com/Wecko-ai)
- **Source:** [Wecko-ai/claude-skills](https://github.com/Wecko-ai/claude-skills)
- **License:** MIT

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-wecko-ai-claude-skills-closing
- Seller: https://agentstack.voostack.com/s/wecko-ai
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.