Handoff
Make any repo AI native — fewer skills, conventions over steps, so any agent can cold-start, implement, self-verify, and write its decisions back.
npx -y skills add initxy/initxy-skills --skill handoffAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 4 stars4 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
Compress the current conversation and work progress into a handoff note. Use for handoff, handover, resuming, summarizing progress, or letting a later session continue.
SKILL.md
1.9 KB, as published. Nobody here has run it
Handoff
Write a handoff note so the next session can continue from the current state without re-excavating.
Flow
- Pin the next objective: if the user specifies what the next session should do, put it first; otherwise organize around the current unfinished tasks.
- Gather state: lay out the goal, changes already made, key decisions, verification results, open items, blockers, and risks.
- Reference, don't copy: for existing implementation notes, ADRs, commits, diffs, docs, and reports, give only the path or URL — don't repeat the full text.
- Protect sensitive info: never write API keys, passwords, tokens, or PII; when needed, just note "credentials must be re-fetched."
- Give the next steps: list the suggested skills, priority actions, files to read first, and things to watch out for.
- Save location: when the user doesn't specify a location, write the handoff note to the system temp directory and give the absolute path in your reply; don't default to writing into the current repo.
Handoff note format
# Handoff
## Next Objective
## Current State
## Completed
## Key Decisions
## Important Files / Links
## Validation
## Open Items
## Risks / Blockers
## Suggested Skills
## Suggested Next Steps
Completion criteria
- The next agent can understand the current goal, state, and best next step.
- Completed and unfinished content are separated, not mixed.
- Key decisions state their reasons and reference the relevant
CONTEXT.md, ADR, implementation note, or code location. - No large existing artifact is copied wholesale; only paths, URLs, or commits are referenced.
- No sensitive information is leaked.