Agent Session Handoff Brief
Turn a coding conversation into a compact handoff with evidence, decisions, blockers, and ordered next steps.
Before & after
Same prompt, same agent — the only difference is the skill.
I can help you summarize the conversation and pass it along to another agent. I'd compile the key decisions, what's been done so far, what's still pending, and anything important they need to know. I could also list any files or documents that are relevant. The handoff document would be structured to make it easy for the next agent to jump in. Do you want me to start drafting a summary, or do you have specific context that's most important to pass along?
HANDOFF DOCUMENT FROM: Claude (Design & Implementation Phase) TO: Claude (Deployment & QA Phase) DATE: 2026-07-31 PROJECT: Event Automation Feature (TikTok Integration) COMPLETED: ✓ Architecture design (ADR-042: Event Streaming via Webhook; /path/to/adr-042.md) ✓ Database schema for event_webhooks table (/prisma/schema.prisma; lines 412-428) ✓ TikTok OAuth flow implemented (/app/api/auth/tiktok/callback.ts) ✓ Webhook signature verification function (/lib/webhook-verify.ts) ✓ Unit tests (42 test cases, 94% coverage; /app/__tests__/webhook.test.ts) PENDING (Assigned to Next Agent): • Integration test: Simulate TikTok webhook payload; verify event_webhooks table populates correctly • Staging deployment: Deploy to staging.agentvolt.app; verify OAuth redirect URLs work • E2E test: Complete flow (user connects TikTok → webhook fires → event stored → user sees event in UI) • Performance test: Load test with 1K webhook events/min; ensure latency stays under 200ms • Security audit: Request penetration test on webhook endpoint before production CRITICAL DECISION LOG: • Chose HTTP 202 (async acceptance) over 200 for webhook responses → allows fault tolerance • Event deduplication key: tiktok_event_id + timestamp (prevents duplicate processing) • Webhook retry policy: 3 exponential backoffs, max 1 hour between attempts KNOWN ISSUES: • TikTok sandbox environment is unstable (50% webhook delivery rate). Workaround: Use production credentials for testing (secure tunnel via ngrok). • Event_webhooks table growth: ~8GB per month at current scale. Add retention policy (archive events >90 days old) before production. NEXT AGENT CONTEXT: • Your focus: Deployment readiness. Block if any integration test fails. • Time estimate: 6 days (integration + E2E testing + perf validation) • Escalation: If load test fails, don't reimplement. Ping the original implementer. • Success criteria: All integration tests pass. E2E test completes in under 300ms. Security audit has zero critical findings. ARTIFACT REFERENCES (by path, not duplicated): • /app/api/webhooks/tiktok/route.ts (webhook handler) • /prisma/schema.prisma (data model) • /app/__tests__/webhook.test.ts (test suite) • /docs/ADR-042-event-streaming.md (architecture rationale) HANDOFF COMPLETE. Next agent has all context to proceed without backfill.
About this skill
Starting a fresh agent should not mean explaining the task again or watching it repeat an abandoned approach. This skill turns the conversation into a compact continuation brief that preserves the current objective, actual work state, decisive context, and the next useful action.
What it does
- Separates completed changes, uncommitted work, failed attempts, and unverified claims.
- Links existing plans, issues, diffs, and decisions using exact paths or URLs.
- Preserves corrections and explains which earlier assumptions they supersede.
- Records test scope and results without turning a partial check into a completion claim.
- Orders remaining work by dependency and gives the next agent a concrete starting point.
- Carries forward user constraints, approval boundaries, and unresolved external operations.
What you get
A paste-ready Markdown handoff sized to the task, with blockers near the top and enough context for another agent to continue without reading the full transcript. When requested and supported, it can be saved to a specified project file.
Best for
Developers moving a long coding session to a fresh context, switching agents during implementation, or handing unfinished engineering work to a teammate.
SKILL.md preview
---
name: agent-session-handoff-brief
description: >-
Compact an active coding conversation into an evidence-grounded continuation brief, separating current state from historical attempts and linking authoritative artifacts instead of copying them. Use when the user says "hand this off", "summarize for a fresh agent", "compact this conversation", "give me a new-session prompt", "another agent needs to pick this up", "the next agent keeps repeating finished work", or "we lost track of what is actually done".
---
# Agent Session Handoff Brief
… (sign up to view the full skill)More development skills
View all Development skills →Performance Bottleneck Profiling: Node, Python, Go
Classifies a slowdown, gives exact profiler commands, reads the output, and proves the fix with same-conditions p95/p99 numbers.
Feature Scaffolding with Working Tests
Turn a short feature spec into code, runnable tests, and integration wiring that match your existing project.
Feature Flag Rollouts and Kill Switches
Design staged flag rollouts, verify rollback paths, and audit flag debt with explicit defaults, gates, and retirement plans.
Microcontroller Firmware and RTOS Engineering
Build and debug MCU firmware with bounded ISR/RTOS timing, safe DMA ownership, and measured power budgets.