handoff
Compact the current conversation into a handoff document for another agent to pick up.
- License 许可
- MIT
- 最近更新
- 2026-09-22
- Stars
- ★ 0
- Forks
- 0
01 前言 Front Matter
- name
- handoff
- description
- Compact the current conversation into a handoff document for another agent to pick up.
- argument-hint
- What will the next session be used for?
- disable-model-invocation
- true
02 资源文件 1
- 参考文档agents/openai.yaml141 B
03 SKILL.md 正文
Write a handoff document summarising the current conversation so a fresh agent can continue the work. Save to the temporary directory of the user's OS - not the current workspace.
Include a "suggested skills" section in the document, which suggests skills that the agent should invoke.
Do not duplicate content already captured in other artifacts (specs, plans, ADRs, issues, commits, diffs). Reference them by path or URL instead.
Redact any sensitive information, such as API keys, passwords, or personally identifiable information.
If the user passed arguments, treat them as a description of what the next session will focus on and tailor the doc accordingly.
04 同仓库更多 Skills
test-driven-development
Drive one observable software behavior at a time through a valid Test-Driven Development cycle: select the test, prove Red for the intended reason, implement the smallest Green change, and preserve the test as regression evidence. Use before adding production behavior, fixing defects, changing contracts, or safely characterizing legacy behavior. Also use when tests are brittle, over-mocked, skipped, weakened, or passing without proving the requirement.
suite-orchestrator
Route complex work across the complete Agent Skill Suite while enforcing content preservation, explicit scope, correct specialist timing, evidence-backed completion, and durable state. Use when a task involves coding, agent operations, repository bootstrapping, conversation compilation, artifact generation, expert review, persistent memory, or several of these together. Also use when the correct skill is unclear or work must pass through multiple specialist skills in a controlled sequence.
solid-design
Design, refactor, and review object-oriented or modular code with the SOLID principles while preserving behavior and avoiding needless abstraction. Use when defining responsibilities, extension seams, subtype contracts, consumer interfaces, dependency direction, modules, services, strategies, plugins, adapters, repositories, or architecture boundaries; also use when test friction, tangled dependencies, large classes, unsupported methods, framework leakage, or repeated variant branches signal a structural problem.
quality-gates
Verify engineering work against required behavioral, architectural, regression, build, security, compatibility, documentation, and diff-hygiene evidence before completion. Use after a code change, defect fix, refactor, migration, generated-code update, or implementation plan appears complete; also use for read-only code reviews and release-readiness checks. Return PASS or BLOCKED without implementing the remediation.
loop-engineering
Execute non-trivial software work through small, observable build-test-observe-adjust loops with explicit hypotheses, evidence, outcomes, and next actions. Use throughout feature implementation, debugging, refactoring, migrations, performance work, and complex verification where feedback must guide each subsequent step. Use when work is stalling, changes are growing too large, failures repeat, or an engineering ledger is required.
local-memory
Maintain fast, durable, curated local memory for AI agents on Android, Termux, Linux, macOS, Windows, or other lightweight environments. Use when an agent must remember stable preferences, project facts, decisions, procedures, failure resolutions, and task state across sessions; retrieve relevant memories before planning; resolve conflicting or stale memories; or provide explicit remember, search, pin, update, forget, status, and export controls without a cloud dependency.