Skip to content
Skills Library
AI & agents · Automation

checkpoint

/checkpoint

Close a session once: daily log, fact reconciliation (master files, project docs, memory), selective commit/push, and remote verification. Use for /checkpoint, end of session, save, push, commit and push. Full: full reconciliation, knowledge sync, memory cleanup. Lite mode: inline, no subagent, for sessions already kept in sync.

Licence
MIT
Last repository change
25.10.2026

Assessment and security

100/ 100Class: good
Format25 / 25
Frontmatter with name and description, valid name, description up to 1,024 characters, name matching the folder.
Discoverability25 / 25
Description explains when to use the skill, has a reasonable length and uses third person.
Content35 / 35
At least 150 words and at most 500 lines, headings, step lists, examples or code.
Hygiene15 / 15
No placeholders such as TODO, no Windows paths, with a licence.
Riskno signals detected in the scan
Files in this package
8
Scripts
2
HTML comments
0
URLs in SKILL.md
0
Deprecation

No signals detected. Repository changed on 25.10.2026, without references to retired models or APIs.

Licence

MIT, open: use and modify while retaining author attribution.

What to check before installing
  • 2 scripts in this package. Review them before installing, as Claude may execute them.

Scripts: scripts/enumerate.sh, scripts/stage-own-edits.sh

What it does and when to use it

Close a session once: daily log, fact reconciliation (master files, project docs, memory), selective commit/push, and remote verification. Use for /checkpoint, end of session, save, push, commit and push. Full: full reconciliation, knowledge sync, memory cleanup. Lite mode: inline, no subagent, for sessions already kept in sync.

Original author description

Close a session once: daily log, fact reconciliation (master files, project docs, memory), selective commit/push, and remote verification. Use for /checkpoint, end of session, save, push, commit and push. Full: full reconciliation, knowledge sync, memory cleanup. Lite mode: inline, no subagent, for sessions already kept in sync.

How it works

  1. Modes
  2. 0. Lite mode
  3. 1. The main thread builds the closure manifest exactly once
  4. 2. Execute exactly once
  5. 3. Main-thread verification
  6. Condensed report
Original SKILL.md headings · 81 lines, 1,024 words
SKILL.md preview
---
name: checkpoint
description: "Close a session once: daily log, fact reconciliation (master files, project docs, memory), selective commit/push, and remote verification. Use for /checkpoint, end of session, save, push, commit and push. Full: full reconciliation, knowledge sync, memory cleanup. Lite mode: inline, no subagent, for sessions already kept in sync."
allowed-tools: Bash(git status *) Bash(git diff *) Bash(git log *) Agent Read
---

# Checkpoint — the single closure entry point

`checkpoint` is the single closure entry point. Its fact-reconciliation engine lives in `references/sync-protocol.md` (neat-freak was merged into this skill on 2026-10-25).

## Modes

- `scoped` (default): write the daily log once; run a targeted reconciliation only if there's a fact list, then commit + push.
- `full`: the user typed `/checkpoint full`, or asked for a full reconciliation, knowledge sync, memory cleanup, or a full closure (formerly `/neat-freak`) — run the complete GATE reconciliation before commit + push.
- `lite`: the user asks for `/checkpoint lite` or lite mode. **The main thread finishes it directly — no subagent, no manifest.** See §0 below.
- `git-only`: only when the user explicitly says no log / no sync.

## 0. Lite mode

**When to use it**: mid-session, you've already been syncing changes into your notes as you went, the daily log is mostly written, and most changes are already committed and pushed. Running the full flow at that point just repeats work that's already done, plus adds another subagent round-trip.

**Trigger** (any one of these is enough):
- The user explicitly asks for lite mode.
- **Automatic downgrade**: the fact list is empty, and 5 or fewer files are attributable to this session. Delegating a closure like that spends a whole subagent round-trip redoing something the main thread could finish in a handful of commands.
- All three hold at once: (1) your notes were updated as you went, not batched at the end; (2) the daily log is already written in-conversation and needs at most a line or two more; (3) there's no cross-file fact reconciliation to do.

Fall back to `scoped` if there's a fact list, more than 5 attributable files, or a mixed-author file that needs a judgment call beyond what `stage-own-edits.sh` can make on its own.

**How to do it** (main thread, inline, no subagent at any point):
1. `git status --porcelain` + `git log -1` across every repo this session touched.
2. Fill in whatever's still missing from the daily log. Don't rewrite what's already there, and don't backfill entries.
3. **Stage only the files attributable to this session** — per the `write-manifest` hook's record (see `hooks/write-manifest/`) — listed one by one, never `git add -A`. For a file that was already modified before this session touched it, or that another session's manifest also lists, don't `git add` the whole file: run `scripts/stage-own-edits.sh <session-id> <file>` instead, so only this session's edits get staged (an exit code of `2` or `3` means hold it and report, don't force it through).
4. Commit + push, and verify `HEAD == origin/<branch>` per repo.
5. Report using the condensed format below, with `sync:` set to `lite (inline)`.

**What lite skips**: the closure manifest, coverage classification, fact-list reconciliation, memory writes, and subagent delegation/verification. Needing any one of those means it isn't lite — fall back to `scoped`.

## 1. The main thread builds the closure manifest exactly once

Use the current platform's task/plan mechanism to build one closure plan, and only one.

1. Get the date and timezone with `date`.
2. Check `git status --porcelain`, `git diff --stat HEAD`, and the current branch for the current repo (and any other project roots listed in the manifest).
3. Review the whole conversation once, producing:
   - `session_summary`: 0–3 items each of what was completed / decided / any pitfalls hit.
   - `fact_list`: facts that changed → the files they affect; not just a list of touched files.
   - `daily_entries`: the condensed content to write into the daily log.
   - `todo_closeouts`: todos in your master files that this session finished or moved forward, each with what to change (mark done with today's date, or rewrite the remaining step). Put anything you're unsure about in warnings for the user instead.
   - `repos`: repo, branch, attributable files, tests run.
   - `coverage`: every changed file, and every external action / scheduled item / payment / todo change, must be tagged `fact_delta: yes` (linked to the fact list) or `fact_delta: no` (with a one-line reason). Anything left unclassified blocks delegation.
4. Compare against today's daily log, recent commits, and the current file state; if another session already wrote the same content more recently, mark `daily_action: skip_duplicate` — don't duplicate or backfill it.
5. The manifest must include `mode`, `memory_mode`, the folder for whichever process scripts are currently in use, the memory folder, the knowledge base root, and every project folder being closed out.

`memory_mode` defaults to `read-only`; it's `update` in `full` mode, or when the user explicitly asked for memory sync or memory cleanup.

## 2. Execute exactly once

Hand the manifest to a fresh-context, general-purpose bounded subagent:

- Codex: use `fork_turns: "none"`, prompt carries only the manifest and this file's path.
- Claude Code: use a fresh general-purpose Agent/Task subagent with `model: sonnet`; it won't see the main thread's conversation history but still loads applicable CLAUDE.md / permissions. The delegation prompt carries only the manifest and this file's path.

The first instruction must tell it to read `references/closure-protocol.md` in full, and not spawn any further subagents. If the platform has no subagent support, or the task is a single-repo `git-only`, the main thread follows that protocol directly.

## 3. Main-thread verification

A subagent's report is a claim, not a fact:

1. If sync edits were made, spot-read 1–2 of the files it claims to have changed.
2. Confirm each item appears in the daily log exactly once: grep a key phrase from each `daily_entries` item, since a worker can report the log as written while the write never happened.
3. Per repo, check the reported hash, `git log -1 --oneline`, and `HEAD == origin/<current-branch>`.
4. Recheck `git status --short`: changes attributable to this session should be closed out; unrelated or concurrent changes should be left untouched and listed.

## Condensed report

```text
daily: written | skipped (<reason>)
sync: skipped | scoped | full (<GATE result>)
commit: <repo> <hash> | clean
vault: <hash> | clean | preserved unrelated changes
pushed: <branches>
warnings: none | <items>
```

No force push; no blind `git add -A`; never sweep unattributable changes in just to make things look clean.
Files in this package (8)
  • SKILL.md 7 KB
  • references/agent-paths.md 4 KB
  • references/closure-protocol.md 5 KB
  • references/dev-mode.md 3 KB
  • references/sync-matrix.md 3 KB
  • references/sync-protocol.md 5 KB
  • scripts/enumerate.sh 8 KB
  • scripts/stage-own-edits.sh 3 KB

Versions

RepositoryScoreLicence
This page · agentcrew-academy/harness-starter-kit100MIT
lvlup-sw/exarchos100Apache-2.0
alienfast/claude100MIT
automateyournetwork/netclaw100Apache-2.0
dtiger1889-ops/claude-harness-toolbox100MIT
Sellyew/Workbuddy-workflow100MIT
spencerneutron/Synapse100MIT
Shweta-Mishra-ai/tokenmizer100MIT
htlin222/dotfiles100MIT

Database snapshot: 13.11.2026 · package version 5fb12d607dd3 (SHA-256 digest of the entire package)

Downloads in the last 30 days: 0 · ranking

Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.

Similar skills

  • 100github-sync
    Review the session's git changes, summarize them for approval, then commit and push to GitHub. Trigger when the user runs /github-sync or says "sync to github", "commit and push", "push this session", "wrap up and push", "save this session to git". For commit-only (no push) or…
  • 100create-checkpoint
    Capture the current session state as a checkpoint artifact per spec section 4.4. Use when the user signals session end or explicitly requests a checkpoint. Do not use mid-session or without user approval.
  • 100sync-push
    Use when the user wants to back up, save, or push their current Claude Code config to GitHub.
  • 100release-checkpoint
    Inspect or create a high-integrity Git commit/release checkpoint with exact repository provenance and staged-diff verification. Use for checkpoint, commit, push-readiness, or release preparation; writes require explicit authorization.
  • 100daily-log
    Daily log conventions for agent memory between sessions. Use when: "daily log", "today's note", "what happened today", "session start", "check context".
  • 100dump
    Session exit and handover -- commit and push work, release claimed PKB tasks, and emit a final report. Use when completing, pausing, or handing off a session.

Feedback and reports

Your message goes to Bartosz Sokoliński. An email address is needed only for a reply. See the privacy policy.