brainstorming-extensio…
Enforces design review gate after brainstorming — bridges superpowers:brainstorming into the metaswarm quality pipeline
Analyze the current session and write a self-contained handoff document so a fresh agent can resume the work with full context — outputs a single "Read XXX.md and do YYY." sentence
$ npx -y skills add dsifry/metaswarm --skill handoff --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/handoffContext preview
The summary Claude sees to decide when to auto-load this skill.
Analyze the current session and write a self-contained handoff document so a fresh agent can resume the work with full context — outputs a single "Read XXX.md and do YYY." sentence
name: handoff description: Analyze the current session and write a self-contained handoff document so a fresh agent can resume the work with full context — outputs a single "Read XXX.md and do YYY." sentence
Capture everything a **fresh agent with zero prior context** needs to resume the current work, write it to a single self-contained markdown document, and hand it off with one unambiguous sentence:
> **Read `docs/handoffs/<file>.md` and do `<the next concrete action>`.**
Use this when a session is ending, when context is about to be compacted, when you are switching machines or sessions, or when the user explicitly asks for a handoff. The goal is **zero context loss**: the receiving agent should be able to act correctly after reading exactly one file.
---
This skill produces exactly two things:
1. **A handoff document** at `docs/handoffs/handoff-<YYYY-MM-DD-HHmm>.md` (created if the directory does not exist). It is comprehensive, self-contained, and links to — or quotes — every artifact the next agent must read. 2. **A single closing sentence**, and nothing else after the document is written, of the exact form:
Read docs/handoffs/handoff-2026-06-17-1432.md and do <YYY>.
Where `<YYY>` is one concrete, actionable next step (an imperative, not a vague theme). Good: "implement the `parseConfig` validation in `src/config.ts:84` and make `tests/config.test.ts` pass". Bad: "continue the work" / "look into the config stuff".
The sentence is the deliverable the user hands to the next agent. The document is what makes that sentence safe to follow.
---
Review the session to date and answer, concretely:
Pull from: the user's original request, any GitHub Issue (`gh issue view <n>`), the design/plan docs, and the arc of the conversation.
If this project uses metaswarm's context persistence, read and fold in whatever exists — do **not** duplicate it blindly, summarize and link:
ls .beads/plans/active-plan.md # the approved plan (if mid-execution) ls .beads/context/project-context.md # completed work units, patterns, tooling ls .beads/context/execution-state.md # current work unit, phase, retry count bd prime --work-type recovery 2>/dev/null # reload plan + state + knowledge, if beads present
If an active plan exists, the handoff must point to it explicitly and state which work unit/phase is in progress.
Be honest and specific about state. Distinguish clearly between:
Capture the working tree reality so the next agent is not surprised:
git branch --show-current git status --short git log --oneline -10 git diff --stat
Note uncommitted changes, stashes, and whether the branch is pushed.
List every artifact the next agent must read **before acting**, and for each one say *why* and *what to look for*. Categories to sweep:
Prefer precise pointers (`src/foo.ts:120-145`) over whole-file references. If something is short and load-bearing, quote it directly into the handoff so the next agent does not have to hunt.
Pick the one most important, concrete next step. It must be:
Create `docs/handoffs/handoff-<YYYY-MM-DD-HHmm>.md` using the template below. Fill every section; write "None" where a section genuinely does not apply rather than deleting it.
After writing the file, verify it (Step 8), then output the single sentence — exactly one line, the literal file path, and the concrete action. Do not add commentary after it.
Confirm, as if you were the receiving agent who knows nothing:
---
# Handoff: <short title of the work> **Date**: <YYYY-MM-DD HH:mm> · **Branch**: `<branch>` · **Author session**: <model/agent> ## 1. Objective <1–2 sentences: what we are trying
A self-improving multi-agent orchestration framework for Claude Code, Gemini CLI, and Codex CLI.
Enforces design review gate after brainstorming — bridges superpowers:brainstorming into the metaswarm quality pipeline
Create comprehensive GitHub issues with TDD plans, acceptance criteria, and agent instructions for autonomous PR lifecycle…
Automatic review gate that runs after brainstorming completes - spawns PM, Architect, Designer, Security, and CTO agents in…
Delegate implementation and review tasks to external AI CLI tools (Codex, Gemini) with cross-model adversarial review
Address PR review feedback systematically — fetch inline comments, review bodies, handle outside-diff-range comments, resolve…
Migrate from npm-installed metaswarm to the marketplace plugin — removes redundant files with safety checks