handoff
Structured work handoff across context boundaries: triage decisions, constraints, and remaining work into a prompt-quality transfer document. Invoke whenever task involves handing off work — session restart, teammate delegation, async resumption, or any context boundary crossing.
What this skill does
# Handoff A handoff document is a prompt — an instruction set for the receiving agent. Triage your context into a structured, information-dense transfer document that preserves high-value state and drops noise. ## Mode Two modes, determined by who receives the handoff: - **Self** — receiver shares project knowledge. Task state, decisions, remaining work. Minimal codebase orientation. - **Teammate** — receiver may lack project context. Adds codebase orientation, skill/tool recommendations, and convention pointers. If the user doesn't specify, ask. ## Triage Separate your context into two categories before writing anything. <preserve> **Keep** — state that must survive the boundary: - Decisions made and their rationale - Constraints discovered during work - External resource identifiers (URLs, issue IDs, file paths, branch names) - Quality/verification state (what was tested, what passed/failed) - Remaining work in priority order - Blockers and open questions </preserve> <drop> **Drop** — noise in transfer: - Intermediate exploration and search paths - Failed approaches (unless the failure constrains remaining work) - Content already committed to files or git history - Content already captured in upstream artifacts (PRDs, plans, ADRs, issue trackers) — reference by path or ID, don't restate - Information derivable from the codebase (file structure, function signatures) - Raw tool output </drop> ## Document Structure Generate a markdown document. Omit sections with no content — empty headers are noise. ### Self-Handoff ```markdown # Handoff: [task name] ## Context [1-2 sentences: what this work is about and its current state] ## Decisions - [decision]: [rationale] ## Constraints - [constraint]: [why it matters] ## External Resources - [resource type]: [identifier/URL] ## Remaining Work 1. [highest priority next step] 2. [subsequent steps...] ## Verification State - [what was tested/verified and result] - [what remains untested] ## Open Questions - [question]: [context needed to resolve] ``` ### Teammate Handoff Same structure as self-handoff, with these sections added before Decisions: ```markdown ## Codebase Orientation - [key file/directory]: [relevance to this task] ## Recommended Skills & Tools - [skill or tool]: [why it's needed] ## Conventions - [convention]: [where it's documented] ``` ## Compression Target 500-2000 tokens. When work is large: - **Decisions** — keep all. Highest-value content. - **Constraints** — keep all. Losing these causes wrong decisions. - **External resources** — keep identifiers, drop obvious descriptions. - **Remaining work** — top 5-7 items. Collapse minor items into "Also:" line. - **Verification state** — summarize as "X of Y verified, [specific failures]." - **Codebase orientation** (teammate mode) — only files the receiver will touch. ## Self-Check After generating the handoff, verify: - [ ] Decisions section is populated (if any decisions were made) - [ ] Remaining work is populated (if work is incomplete) - [ ] External resource IDs are present (if external systems were used) - [ ] No raw tool output or intermediate exploration leaked in - [ ] Document is under 2000 tokens If Decisions or Remaining Work is empty when work was done, the triage missed something. Re-examine the conversation. ## Critical Rules - **Print, don't write.** Default to printing the handoff to the conversation so the user can copy it. Do not save to a file unless the user explicitly asks for one (e.g. "write it to handoff.md"). The user controls delivery. - **No information gathering.** Read only from context already in the conversation. Do not fetch files, search code, or query memory — context may be nearly full. - **Rationale with every decision.** A decision without rationale forces re-derivation or blind trust. - **Identifiers, not descriptions.** "PR #247" not "the pull request we opened." "src/auth/middleware.ts:42" not "the auth file around line 42." - **Redact secrets and PII.** Strip API keys, tokens, passwords, and personal data before emitting — the document gets printed, copied, and shared. - **The document is standalone.** The receiver has no access to this conversation.
Related in AI Agents
skill-development
IncludedComprehensive meta-skill for creating, managing, validating, auditing, and distributing Claude Code skills and slash commands (unified in v2.1.3+). Provides skill templates, creation workflows, validation patterns, audit checklists, naming conventions, YAML frontmatter guidance, progressive disclosure examples, and best practices lookup. Use when creating new skills, validating existing skills, auditing skill quality, understanding skill architecture, needing skill templates, learning about YAML frontmatter requirements, progressive disclosure patterns, tool restrictions (allowed-tools), skill composition, skill naming conventions, troubleshooting skill activation issues, creating custom slash commands, configuring command frontmatter, using command arguments ($ARGUMENTS, $1, $2), bash execution in commands, file references in commands, command namespacing, plugin commands, MCP slash commands, Skill tool configuration, or deciding between skills vs slash commands. Delegates to docs-management skill for official documentation.
reprompter
IncludedTransform messy prompts into well-structured, effective prompts — single or multi-agent. Use when: "reprompt", "reprompt this", "clean up this prompt", "structure my prompt", rough text needing XML tags and best practices, "reprompter teams", "repromptception", "run with quality", "smart run", "smart agents", multi-agent tasks, audits, parallel work, anything going to agent teams. Don't use when: simple Q&A, pure chat, immediate execution-only tasks. See "Don't Use When" section for details. Outputs: Structured XML/Markdown prompt, quality score (before/after), optional team brief + per-agent sub-prompts, agent team output files. Success criteria: Single mode quality score ≥ 7/10; Repromptception per-agent prompt quality score 8+/10; all required sections present, actionable and specific.
adaptive-compaction
IncludedAdaptive add-on policy and recovery layer that decides WHEN to compact, prune, snapshot, or fork -- replacing fixed-percent auto-compaction across Claude Code, Codex, and MCP-capable hosts. Trigger on auto-compact timing or damage: "when should I compact", "is it safe to compact now or start a fresh session", "auto-compact fires too early/mid-task", "switching to an unrelated task but the window still has space", "context rot", "answers get worse the longer the session runs", "the agent forgot the plan or my decisions after it summarized", "add a layer on top that manages context without changing the agent", raising autoCompactWindow to give the policy room, or installing/tuning a cross-tool compaction policy or PreCompact hook -- even when "compaction" is never said but the problem is context-window pressure or post-summarization memory loss. Do NOT use to summarize a conversation, build RAG, write a summarization prompt (decides WHEN not HOW), or answer max-context-length trivia.
agent-skill-creator
IncludedCreate cross-platform agent skills from workflow descriptions. Activates when users ask to create an agent, automate a repetitive workflow, create a custom skill, or need advanced agent creation. Triggers on phrases like create agent for, automate workflow, create skill for, every day I have to, daily I need to, turn process into agent, need to automate, create a cross-platform skill, validate this skill, export this skill, migrate this skill. Supports single skills, multi-agent suites, transcript processing, template-based creation, interactive configuration, cross-platform export, and spec validation.
llm-wiki
IncludedUse when building or maintaining a persistent personal knowledge base (second brain) in Obsidian where an LLM incrementally ingests sources, updates entity/concept pages, maintains cross-references, and keeps a synthesis current. Triggers include "second brain", "Obsidian wiki", "personal knowledge management", "ingest this paper/article/book", "build a research wiki", "compound knowledge", "Memex", or whenever the user wants knowledge to accumulate across sessions instead of being re-derived by RAG on every query.
skill-master
IncludedAgent Skills authoring, evaluation, and optimization. Create, edit, validate, benchmark, and improve skills following the agentskills.io specification. Use when designing SKILL.md files, structuring skill folders (references, scripts, assets), ingesting external documentation into skills, running trigger evals, benchmarking skill quality, optimizing descriptions, or performing blind A/B comparisons. Keywords: agentskills.io, SKILL.md, skill authoring, eval, benchmark, trigger optimization.