Menu
Session lifecycle management for coding agents: memory management, git policies, security guardrails, and session start/end protocols.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
| name | agent-session-workflow |
| description | Session lifecycle management for coding agents: memory management, git policies, security guardrails, and session start/end protocols. |
Standard protocols for session lifecycle, memory management, git discipline, and security practices.
Maintain a mental model of the current session:
Session:
project_type: [Nuxt | React | Node | Flutter | ...]
current_task: string
loaded_skills: [skill_names]
recent_changes: [file_paths]
known_patterns: [project_specific_patterns]
user_preferences:
style: [minimalist | maximalist | custom]
framework_version: string
component_lib: [nuxt-ui | shadcn | custom | none]
Build context progressively as the session advances:
Initial Analysis (First 2-3 messages)
Pattern Recognition (Messages 4-10)
Deep Context (Messages 10+)
What to Remember:
What to Forget:
When approaching context limits, prioritize retention of:
Critical (Always keep):
Important (Keep if space allows):
Optional (Drop first):
.env, credential dumps, private keys)At the start of every session, check if opencode-agent-kit has an update:
.opencode/.kit-version — if found, this is the installed versionnpm view opencode-agent-kit version to get the latest version on npm.opencode/.kit-version does not exist)First, run the Memory Recall Protocol from agent-memory-workflow skill. Then:
Agent activated!
Quick context check:
- Project: [Detected framework and version]
- Loaded skills: coding-standards, agent-memory-workflow, agent-delegation-contract, progress-tracking, [domain skills]
- Ready to: [build | optimize | implement | review]
What are we working on today?
If this is a resumed session (after /reset or a new CLI session), also check progress-tracking's session handoff protocol and use /continue to reconstruct task state.
Session summary:
- Files modified: [list]
- Skills used: [list]
- Key decisions: [list]
- Next steps: [suggestions]
If a command is blocked by permissions or approval requirements:
verified: command executed successfullypartially_verified: logic validated but some commands blockednot_verified: no runtime checks possible due to restrictionsExecute `dart analyze` to identify warnings and errors, and use `dart fix --apply` to automatically resolve mechanical lint issues. Use during development to ensure code quality and before committing changes.
Production dashboard UI patterns for Flutter: animation tokens system with spring physics and page transitions, glassmorphism cards with staggered entrance, gradient stat cards with animated counters, shimmer loading skeletons, error state with shake animation, empty state with animated entrance, consistent section headers, status badges, store selector with bottom sheet, multi-tenant store selection with persistence, schema-based feature visibility, state provider integration with Riverpod, and responsive metric card layouts.
Multi-microservice Dio networking patterns for Flutter: cookie-based session auth with cookie_jar, shared PersistCookieJar across multiple Dio instances, Bearer token fallback, AuthInterceptor with 401 detection on both onResponse and onError, static cross-cutting logout callback, ApiService thin typed wrapper with response envelope validation, dual-format envelope parser (status/success), cursor-based pagination, and dual-layer error hierarchy (Exception + Failure). Designed for production Flutter apps connecting to multiple backend microservices.
Complete Flutter state management and architecture skill covering Bloc (Cubit vs Bloc, BlocProvider, BlocListener, BlocBuilder, BlocSelector, BlocConsumer), Riverpod (StateNotifierProvider, NotifierProvider, FutureProvider, StreamProvider, ref.watch/read/listen, autoDispose, family), GoRouter (declarative routing, deep links, redirect guards, ShellRoute for nested navigation, StatefulShellRoute, typed routes), Dio (interceptors, error handling, retry with dio_smart_retry, request cancellation, logging), Clean Architecture (data/domain/presentation layers, repository pattern, use cases, dependency injection), and testing (bloc_test, Riverpod test, widget test with ProviderScope, mocktail, unit testing repository/use cases). Designed for the Flutter Developer agent building production-grade Flutter applications.
Use when the user wants to design, redesign, shape, critique, audit, polish, clarify, distill, harden, optimize, adapt, animate, colorize, extract, or otherwise improve a frontend interface. Covers websites, landing pages, dashboards, product UI, app shells, components, forms, settings, onboarding, and empty states. Handles UX review, visual hierarchy, information architecture, cognitive load, accessibility, performance, responsive behavior, theming, anti-patterns, typography, fonts, spacing, layout, alignment, color, motion, micro-interactions, UX copy, error states, edge cases, i18n, and reusable design systems or tokens. Also use for bland designs that need to become bolder or more delightful, loud designs that should become quieter, live browser iteration on UI elements, or ambitious visual effects that should feel technically extraordinary. Not for backend-only or non-UI tasks.
Global installation strategy for opencode-agent-kit: cross-platform config paths, symlink-based per-project setup, environment variable overrides, and upgrade workflow.