Skip to main content
Agent skill

Session Wrapup

End-of-session checkpoint — append diary, update active sprint doc with interim results/retro, write tech-lead handoff, commit and push. Use before /compact or at end-of-day. Distinct from sprint-wrap-up which closes a sprint.

Get started

Get the complete skill folder, including its scripts and reference files.

When to use this skill

Use Session Wrapup when an AI agent needs a reusable SKILL.md workflow for this job: End-of-session checkpoint — append diary, update active sprint doc with interim results/retro, write tech-lead handoff, commit and push. Use before /compact or at end-of-day. Distinct from sprint-wrap-up which closes a sprint.

When not to use it

Skip Session Wrapup when the task is outside the coding category, or when a more specific skill in this directory already covers the same workflow with clearer triggers.

How to install

Where will you use it?

Install scope
Claude Code folder setup
mkdir -p ~/.claude/skills/session-wrapup
# Save SKILL.md to ~/.claude/skills/session-wrapup/SKILL.md
  1. Claude Code personal install: place the skill and its bundled files at ~/.claude/skills/session-wrapup/SKILL.md. For a shared project, commit the folder at .claude/skills/session-wrapup/ instead.
  2. Current Claude Code detects SKILL.md changes in watched directories during the session. If you just created a top-level skills directory, run /reload-skills. Check that the skill is listed before testing its trigger.
  3. For Claude chat or Cowork in the Desktop app, manage uploaded skills in Customize > Skills; copying files into a local Claude Code folder is not a Desktop install. Other agents may use different install locations.

Full install guide for Claude, Cursor, and Codex →

Try a starter prompt

After installing the skill, paste this starter prompt into your agent to verify it triggers on a small task.

Starter prompt
Use the Session Wrapup skill for this task: At end-of-day / session end (user signals "we're done for today"). Confirm any required inputs or files before making changes.

What supports this recommendation?

82/100 provisional content score. This measures documentation, practical detail and source signals. It is not a measured task success rate or a security certification.

Documentation
92/100
Practical detail
92/100
Source evidence
58/100
Source signals
72/100
Model compatibility evidence
  • claudeinferred

    The skill text mentions Claude or a closely associated term. This is keyword-derived compatibility; no hands-on model test is recorded.

  • chatgptuntested

    No model-specific signal or recorded compatibility test was found.

  • geminiuntested

    No model-specific signal or recorded compatibility test was found.

  • copilotuntested

    No model-specific signal or recorded compatibility test was found.

  • llamauntested

    No model-specific signal or recorded compatibility test was found.

  • perplexityuntested

    No model-specific signal or recorded compatibility test was found.

  • mistraluntested

    No model-specific signal or recorded compatibility test was found.

  • grokuntested

    No model-specific signal or recorded compatibility test was found.

Inferred compatibility is based on content signals. Check the app, its available tools and setup requirements before running a skill.

What this skill does

# Session Wrap-Up

Persist all session-level learnings and state to disk BEFORE running `/compact` or ending a session. Ensures nothing is discarded with the conversation.

## When to use

- At end-of-day / session end (user signals "we're done for today") - Before running `/compact` (especially mid-sprint) - After a long merge wave or planning session - When approaching a token budget threshold (≥40% weekly)

**Distinct from `sprint-wrap-up`** which is for closing a sprint with retrospective + tagging. This skill is for routine session ends while a sprint is still active.

## Step 1: Verify on main in /workspace

```bash pwd && git branch --show-current ```

Must be `/workspace` on `main`. Abort if not.

## Step 2: Append to plan/diary.md

Add a new `## YYYY-MM-DD HH:MM — <brief title>` section at the bottom with:

- **Pass rate** — start and end numbers if changed - **Merges** — PRs that landed, with net deltas - **Closes** — PRs rejected, with reasons - **New issues filed** — numbers + one-line purpose - **Incidents** — OOM, stuck processes, false-positive regressions, CI glitches - **Context/budget** — tokens burned, triggers, mitigations applied - **Key learnings** — non-obvious findings worth saving (1-sentence each)

Keep it to ~30-60 lines. Bullets, not prose.

## Step 3: Update plan/issues/sprints/N/sprint.md

Fill the `## Results` section (mark as "interim" if sprint still active):

```markdown ## Results (interim — YYYY-MM-DD, sprint still active)

**Baseline progress:** <start> → **<end>** pass / <total> total = **<pct>%** (<+delta>) Sprint goal (<goal>) <met|not yet met>. <gap> tests <remaining>.

### Merged - **#NNN <title>** (<PR>, <+delta> pass) ...

### Closed without merging - **#NNN <reason>** ... ```

And if the sprint has a `## Retrospective` section, append an interim retro with "What went well / What went badly / Process improvements / Key numbers / Sprint-close criteria remaining".

## Step 4: Append the retrospective section in plan/issues/sprints/N/sprint.md

If a retrospective file doesn't exist yet, create one with frontmatter:

```markdown --- sprint: Sprint-N status: interim | final session_end: YYYY-MM-DD ---

# Sprint N Retrospective — INTERIM (sprint still active) ```

Then fill with the same what-went-well / badly / process-improvements sections. Mark "interim" clearly if sprint isn't closed yet.

## Step 5: Update plan/agent-context/tech-lead.md

```markdown --- agent: tech-lead session_end: YYYY-MM-DD next_session_entry_point: read this file, then plan/issues/sprints/N/sprint.md ---

# Tech Lead Context Summary — YYYY-MM-DD

## Baseline state ## Sprint N status ### Merged today ### Closed ### PRs still in flight ### Unassigned ready work (prioritized) ## Team state ## Local state needing push ## Unfinished work / loose ends ## Entry points for next session ```

This file is what the NEXT session reads as its first tool call instead of `claude --resume`. Be thorough — ~100-200 lines of actionable bullets.

## Step 6: Stage and commit

```bash git add plan/diary.md plan/issues/sprints/N/sprint.md plan/agent-context/tech-lead.md git diff --cached --stat git commit -m "docs(session): end-of-session checkpoint — YYYY-MM-DD

✓

Persist session learnings to disk before /compact. - plan/diary.md: <summary> - plan/issues/sprints/N/sprint.md: <summary + retrospective updates> - plan/agent-context/tech-lead.md: handoff for next session

Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]>" ```

## Step 7: Push

```bash git pull --rebase origin main 2>&1 | tail -3 git push origin main 2>&1 | tail -3 ```

If the rebase gets stuck mid-pick with "editing commit" state, run `git rebase --continue` — it's a transient git state-machine glitch from empty-after-rebase detection.

## Step 8: Tell the user

Short summary: ``` Pre-compact snapshot committed (<commit-hash>). Pushed. Next session: read plan/agent-context/tech-lead.md, do NOT --resume this session. Safe to /compact now. ```

## Notes

- **Never skip step 2 and 3.** Diary + sprint doc are the persistent project log. Compacting without updating them loses the session's learnings permanently. - **tech-lead.md is NOT optional** — it's the handoff channel that replaces session resume. Write it every time you end a long session. - **Don't wait for sprint close** to fill retrospective sections. Interim retros are fine; the sprint doc carries multiple interim entries over time. - **Interim retro files are allowed** — mark `status: interim` in the frontmatter. When the sprint closes, update to `status: final` and amend.

Intended uses

  • At end-of-day / session end (user signals "we're done for today")
  • Before running /compact (especially mid-sprint)
  • After a long merge wave or planning session
  • When approaching a token budget threshold (≥40% weekly)

Implementation notes

  • Review the source instructions and adapt inputs before running the workflow.

Try it, then tell us how it went

A copied prompt does not tell us whether the task worked. Report your result after checking the output. We collect your tool and outcome, never your files or prompt.