Executes these capabilities
Inside this skill
The internal capability tree shows how this skill gets its work done.
SAgent Handoff
- state summarization
- next-step definition
- risk flagging
The contract
Input
- The current session state
- Files changed and decisions made
- Open questions
Output
- A handoff note covering status, decisions, next steps, and risks
Deliverables
- HANDOFF.md or equivalent
Use cases
- Ending a work session
- Switching agents mid-task
- Resuming after a long break
Dependencies and replacements
Dependencies
- project-memory for durable context
Required tools
- A text editor
Required permissions
- Write access to the repository
Compatible with
- project-memory
- validation-checklist
Alternatives
- Verbal handoff with the next agent
- Issue descriptions
Replaceable components
- The file name and template
Reliability
Human checkpoints
- Confirm the handoff captures the true state
Limitations
- A handoff is a snapshot, not a guarantee
- Loses value if not updated at the end of each session
Failure modes
- Handoff written from memory instead of the diff
- Next steps stated as vague intentions
- Risks hidden to avoid admitting problems
Recovery steps
- Regenerate the handoff from the latest git diff
- Rewrite next steps as concrete tasks with acceptance criteria
- List risks without judgment
Validation methods
- A fresh agent follows the handoff to the stated next step
Sources
Entry checked on Jul 31, 2026. Verification status reflects what AgentSkillTree has independently run, not the skill's own claims.
When to use
Use at the end of any session where another agent or another day will continue the work. The handoff answers: where are we, what changed, what is next, and what could go wrong.