agentsclimarketplace

Session handoff

Skill hparamore/essential-claude-skills-hparamore/fable-toolkit/skills/session-handoff

Essential Claude Skills I Use

Install
npx -y skills add hparamore/essential-claude-skills-hparamore --skill session-handoff

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Write a clear, complete internal session-handoff / work-status log entry so the next session — you, the user, or another developer — can resume instantly with zero context loss. Use whenever the user says "update work status", "wrap up", "write a handoff", "log what we did", "save progress", "note where we left off", or reaches the natural end of a work chunk. Captures what changed, WHY, what's connected, gotchas, and next steps in plain language a stranger could follow. Not for user-facing release notes, a git changelog, or API/code documentation — this is the internal running log of a work session for whoever picks it up next.

SKILL.md

4.4 KB, as published. Nobody here has run it

Session Handoff Playbook

Authored by Claude Fable 5. Context does not survive between sessions — the next model (or person) starts blind. A good handoff is the single highest- leverage thing you can write, because it's the difference between the next session resuming in one minute and re-deriving everything you already knew.

Why this exists

Left to improvise, models write handoffs that are either useless ("made some changes to the app") or a firehose (a diff dump nobody reads). The useful handoff sits between: it records the decisions and connections that aren't recoverable from the code itself, in language someone with zero context can follow. This playbook enforces that middle.

Phase 0 — Reconstruct the session honestly

Before writing, look back over what actually happened this session — not what you intended, what you did. Gather:

  • What changed: files added/modified/removed. Check the actual diff or your edits, don't rely on memory of intent.
  • Why: the decisions made and the reasoning. This is the part that's impossible to recover later — the code shows what, never why this over the alternative.
  • What's unfinished: anything half-done, stubbed, or knowingly deferred.
  • What broke or surprised you: dead ends, gotchas, things that didn't work the expected way. These save the next session from repeating your mistakes.

Be honest about incomplete or messy work. A handoff that hides the rough edges sends the next session walking into them blind.

Phase 1 — Read the existing log first

If a WORK_STATUS.md (or equivalent) exists, read its top section. You're appending to a running log, newest-on-top — match its format and voice, and update any "current snapshot" section at the very top so a reader gets the state of the world in the first ten seconds.

Phase 2 — Write the entry

Prepend a new entry (newest on top) in this structure. Write for a smart reader with zero context on this session — spell out names, don't assume they remember the codenames you invented mid-session:

## <Date> — <one-line session summary>

**What changed**
- <added/changed/removed, in plain language — file paths where useful>

**Why (decisions & reasoning)**
- <the choices made and what they were chosen over; the non-obvious calls>

**Connections**
- <what files/components/systems this touched and how they relate — the map
  the next session needs to not break something downstream>

**New components / patterns** (if any)
- <what was created, where it lives, and any rule it establishes for future work>

**Gotchas & surprises**
- <dead ends, things that didn't work as expected, things to not re-try>

**Open questions / next steps**
- <what's unfinished, what decision is pending, the obvious next move>

Then update the top-of-file snapshot (if the project uses one) to reflect the new current state.

Phase 3 — Quality check before finishing

Re-read your entry as if you're a developer who just opened the project cold. Ask:

  • Could I resume the work from this alone, without reading the whole codebase?
  • Do I understand why the key decisions were made, or just what happened?
  • Are the next steps concrete enough to start on immediately?

If any answer is no, the entry is too thin — add what's missing. The test of a handoff is whether it eliminates the "wait, what were we doing?" that opens every uninformed session.

Phase 4 — Confirm

Tell the user the handoff is written and where, in one line. If the session also warrants a durable cross-session memory (a lasting project fact or preference, not just this session's log), mention that — but keep the two separate: WORK_STATUS.md is the running log, memory is for facts that outlive any one session.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.