agentleFS
kev1nzh37kev1nzh37/handoff/skills/handoff-handon/SKILL.md

handoff-handon

Skill0 starsChanged 4 months agoMITView on GitHub

Create and restore repo-local coding handoffs for Codex, GPT, Claude, or other agentic development sessions. Use when the user says `/handoff`, `handoff`, "prepare a handoff", "ready to leave work", `/handon`, `handon`, "continue from last time", "restore context", or asks to preserve or resume project state across machines, sessions, or chat history loss.

  • Reads credentials

The file

Show its rules and text

We split this skill into the rules it gives an agent

21 rules. About 0 are common, 6 are distinctive, and 15 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈6

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 15

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use repo-local markdown files to preserve only the information needed to continue work in a fresh session.

    line 10
  2. Very specific:

    If the user asks for /handoff or is ending a work session, run the Handoff workflow.

    line 25
  3. Very specific:

    If the user asks for /handon or wants to continue from a previous handoff, run the Handon workflow.

    line 26
  4. Very specific:

    If the repo does not yet contain the handoff files and the user asks to set up the workflow, create the structure using …

    line 27
  5. Very specific:

    If the request is ambiguous, infer from intent: leaving/saving state means handoff.

    line 28
  6. Very specific:

    When creating or updating a handoff:

    line 32
  7. Distinctive:

    Read existing context when present:

    line 43
  8. Very specific:

    Create docs/HANDOFF/ if missing.

    line 54
  9. Very specific:

    Update docs/HANDOFF/CURRENT.md using the Current Handoff template in references/templates.md.

    line 55
  10. Very specific:

    Update docs/HANDOFF/TODO.md or docs/HANDOFF/DECISIONS.md only when the user request or discovered project state clearly requires it.

    line 56
  11. Very specific:

    Keep the final response short.

    line 57
  12. Very specific:

    If the user says they will continue on another machine, remind them that /handoff writes repo-local files only.

    line 70
  13. Very specific:

    If docs/HANDOFF/CURRENT.md is missing, say exactly:

    line 94
  14. Very specific:

    If the user expected cross-machine continuity, remind them that the previous machine must commit and push docs/HANDOFF/ or otherwise sync …

    line 107
  15. Very specific:

    If the current branch differs from the branch recorded in CURRENT.md, stop and ask for branch confirmation before modifying files.

    line 109
  16. Distinctive:

    If local uncommitted changes exist, state that handon will summarize first and will not modify files until confirmed.

    line 110
  17. Distinctive:

    Do not modify code or project files during the first handon response.

    line 112
  18. Very specific:

    Never write these into handoff files:

    line 116
  19. Distinctive:

    If suspicious sensitive material is encountered, describe it abstractly:

    line 130
  20. Distinctive:

    Avoid recording sensitive usernames, internal account names, production hostnames, or private filesystem paths when a generic description …

    line 137
  21. Very specific:

    Read references/templates.md when creating missing handoff files or formatting CURRENT.md, DECISIONS.md, TODO.md, AGENTS guidance, or the …

    line 141

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Handoff / Handon

Overview

Use repo-local markdown files to preserve only the information needed to continue work in a fresh session. Do not save full chat history, secrets, or a chronological transcript.

Default project files:

AGENTS.md
docs/HANDOFF/CURRENT.md
docs/HANDOFF/DECISIONS.md
docs/HANDOFF/TODO.md

CURRENT.md is the core handoff artifact. It should let a new machine, new agent, or new context window recover the current objective, recent work, risks, next steps, and validation plan in about three minutes.

Workflow Decision

  • If the user asks for /handoff or is ending a work session, run the Handoff workflow.
  • If the user asks for /handon or wants to continue from a previous handoff, run the Handon workflow.
  • If the repo does not yet contain the handoff files and the user asks to set up the workflow, create the structure using references/templates.md.
  • If the request is ambiguous, infer from intent: leaving/saving state means handoff; returning/restoring state means handon.

Handoff Workflow

When creating or updating a handoff:

  1. Inspect project state before writing:
git status
git diff --stat
git diff
git log --oneline -5
  1. Read existing context when present:
README.md
AGENTS.md
docs/HANDOFF/TODO.md
docs/HANDOFF/DECISIONS.md
docs/HANDOFF/CURRENT.md
  1. Inspect recently changed core files as needed to understand the actual implementation state.
  2. Create docs/HANDOFF/ if missing.
  3. Update docs/HANDOFF/CURRENT.md using the Current Handoff template in references/templates.md.
  4. Update docs/HANDOFF/TODO.md or docs/HANDOFF/DECISIONS.md only when the user request or discovered project state clearly requires it.
  5. Keep the final response short. Do not paste the full handoff unless the user asks.

Required final response shape:

Handoff completed.

Updated:
docs/HANDOFF/CURRENT.md

List additional updated handoff files only if they were actually changed.

If the user says they will continue on another machine, remind them that /handoff writes repo-local files only. They must commit and push docs/HANDOFF/ or otherwise sync those files before running /handon elsewhere.

Handon Workflow

When restoring context:

  1. Read:
AGENTS.md
docs/HANDOFF/CURRENT.md
docs/HANDOFF/DECISIONS.md
docs/HANDOFF/TODO.md
README.md
  1. Check:
git status
git branch --show-current
git log --oneline -5
  1. If docs/HANDOFF/CURRENT.md is missing, say exactly:
No handoff file found.

Expected:
docs/HANDOFF/CURRENT.md

Handon cannot restore the previous project state from handoff.

You may infer context from README, git log, and TODO only after clearly labeling it as inference.

If the user expected cross-machine continuity, remind them that the previous machine must commit and push docs/HANDOFF/ or otherwise sync those files before /handon can restore context here.

  1. If the current branch differs from the branch recorded in CURRENT.md, stop and ask for branch confirmation before modifying files.
  2. If local uncommitted changes exist, state that handon will summarize first and will not modify files until confirmed.
  3. Output a short handon recap using the concise format in references/templates.md.
  4. Do not modify code or project files during the first handon response. Wait until the user explicitly confirms with words such as "start", "continue", "go ahead", or equivalent non-English confirmation phrases.

Safety Rules

Never write these into handoff files:

API keys
tokens
cookies
passwords
private keys
.env actual values
auth.json contents
database credentials
company-sensitive information

If suspicious sensitive material is encountered, describe it abstractly:

- Sensitive credentials were detected during the session, but they were not written into this handoff.
- Check local environment files before sharing or committing.

Avoid recording sensitive usernames, internal account names, production hostnames, or private filesystem paths when a generic description is enough.

References

  • Read references/templates.md when creating missing handoff files or formatting CURRENT.md, DECISIONS.md, TODO.md, AGENTS guidance, or the Handon summary.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.