Verifiable cognition for AI agents — a tamper-evident, replayable ledger of what an agent did and why (korg-ledger spec).
Setup
Choose your client, copy the config, then fill in any required environment variables.
claude mcp add korg-introspect -- npx -y @korgg/introspect-mcp~/.codex/config.toml
[mcp_servers.korg-introspect]
command = "npx"
args = ["-y", "@korgg/introspect-mcp"]
# codex mcp add korg-introspect -- npx -y @korgg/introspect-mcp~/.cursor/mcp.json
{
"mcpServers": {
"korg-introspect": {
"command": "npx",
"args": [
"-y",
"@korgg/introspect-mcp"
]
}
}
}.vscode/mcp.json
{
"servers": {
"korg-introspect": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@korgg/introspect-mcp"
]
}
}
}~/.gemini/settings.json
{
"mcpServers": {
"korg-introspect": {
"command": "npx",
"args": [
"-y",
"@korgg/introspect-mcp"
]
}
}
}claude_desktop_config.json
{
"mcpServers": {
"korg-introspect": {
"command": "npx",
"args": [
"-y",
"@korgg/introspect-mcp"
]
}
}
}.trae/mcp.json
{
"mcpServers": {
"korg-introspect": {
"command": "npx",
"args": [
"-y",
"@korgg/introspect-mcp"
]
}
}
}