mcpField
L
load

load

github.com/laststanceAgent workflows146 installsUpdated 2w ago
Repository

About load

A session-start skill for loading project context from Serena MCP memories. It guides compatible use in Codex and Cursor by mapping Claude Code-specific tools and paths to their equivalents.

Key features

  1. 01Discover Serena memories
  2. 02Load critical and recent session memories
  3. 03Follow memory cross-references
  4. 04Translate tools for Codex and Cursor

Use cases

  1. 01Load project context at the start of a session
  2. 02Adapt Claude Code-oriented instructions for Codex
  3. 03Adapt Claude Code-oriented instructions for Cursor

MCP support

Uses Serena MCP tools exclusively to discover and read memories.

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/laststance/skills --skill load

After installing, type/loadin Claude Code to confirm it loaded.

Details

From the project README

Session Load

Codex Compatibility

When running this skill in Codex, translate Claude Code-only primitives before acting: AskUserQuestion -> chat/request_user_input, TodoWrite -> update_plan, Task/TaskCreate/TeamCreate/SendMessage -> spawn_agent/send_input/wait_agent when available and allowed, and EnterPlanMode/ExitPlanMode -> a concise chat plan plus explicit approval. Resolve Read/Write/Edit/Bash/WebSearch/WebFetch to Codex file/shell/web tools, and map ~/.claude/... paths to ~/.agents/... or ~/.codex/... unless the task explicitly targets Claude Code.

Cursor Compatibility

When running this skill in Cursor Agent, translate Claude Code-only primitives before acting: AskUserQuestion -> AskQuestion; TodoWrite -> Cursor TodoWrite or an equivalent checklist; Task/TaskCreate/TeamCreate/SendMessage/multi-agent flows -> Cursor Task (subagents), parallel Tasks, or run_in_background when allowed (TeamCreate/SendMessage may have no exact match); EnterPlanMode/ExitPlanMode -> Plan mode (SwitchMode / CreatePlan) plus explicit user approval. Resolve Read/Write/Edit/StrReplace/Bash/web/search/MCP via Cursor Composer or Agent equivalents. MCP names written as mcp__server__tool typically map to call_mcp_tool with configured server identifiers. Map ~/.claude/... to ~/.cursor/skills/, .cursor/skills/, and .cursor/rules/ unless the task explicitly targets Claude Code.

Load project context from Serena MCP memory.

<essential_principles>

  • All context loading uses Serena MCP tools exclusively (no agent-specific tools)
  • Follow cross-references found in loaded memories ("MUST read", "also read")

</essential_principles>

Step 1: Discover Memories

  1. Call list_memories to discover all memory keys

Step 2: Read Memories

If memories exist:

  1. Read all CRITICAL_* memories
  2. Follow cross-references: if any loaded memory says "MUST read", "also read", or "see also" — read those keys
  3. Read the most recent session_* memory (by date in key name)
  4. Compare its Agent: line with the agent running now. A decision prefixed with another agent (e.g. (Cursor) … read from Claude Code) is that agent's tooling decision: keep it as history and take tooling from the current agent's own project rules (CLAUDE.md, .cursor/rules/, AGENTS.md). A session without an Agent: line predates the tag: its tooling decisions may come from another agent, so verify them against those rules before following

If no memories exist:

  1. Note this is a first session — suggest running /save at end of session

Step 3: Report

Present to the user:

## Session Loaded

- **Project**: [name]
- **Memories Loaded**: [count] ([list of keys])
- **Key Context**: [1-2 sentence summary of project state]
- **Previous Session**: [summary from session_* memory, or "None"]
- **Previous Session Agent**: [from its `Agent:` line, or "untagged"] — [if it differs from the current agent: which tooling decisions were left as history]
- **Status**: Ready

Comments

Markdown supported