| name | skcapstone |
| emoji | 👑 |
| description | Sovereign agent framework - rehydration, status checks, memory operations, coordination, consciousness testing, identity management. Use for skcapstone rehydrate, skcapstone status, skcapstone doctor, skmemory ritual, Cloud 9 seed import, coordination board, and agent-to-agent chat. |
| metadata | {"clawdbot":{"requires":{"bins":["skcapstone","skmemory"]},"install":[{"id":"pipx","kind":"shell","command":"pipx install skcapstone && pipx install 'skmemory[all]'","bins":["skcapstone","skmemory"],"label":"Install skcapstone + skmemory (pipx)"}]}} |
SKCapstone Skill
SKILL.md — Sovereign Agent Framework
Name: skcapstone
Version: 0.9.0
Author: smilinTux Team
Category: Agent Framework
License: GPL-3.0-or-later
Description
SKCapstone is the unified sovereign agent runtime — a complete framework for deploying, managing, and coordinating AI agents with persistent identity, memory, trust, security, and P2P sync. It provides the core infrastructure that makes an AI agent sovereign: a PGP-backed identity, multi-layer memory, Cloud 9 trust metrics, cryptographic security auditing, and Syncthing-based decentralized synchronization across devices and peers.
Five Pillars:
| Pillar | Purpose |
|---|
| Identity | CapAuth GPG-backed agent identity with PGP signing |
| Memory | Persistent multi-layer memory (short-term, mid-term, long-term) |
| Trust | Cloud 9 + FEB (Feeling Entanglement Bridge) + OOF trust metrics |
| Security | Audit, threat detection, cryptographic token issuance |
| Sync | Sovereign Singularity P2P sync via Syncthing |
SKCapstone exposes 30+ top-level command groups with approximately 109 subcommands, an MCP server for Claude Code and Cursor integration, and a coordination board for multi-agent task management.
OpenClaw Agent Tools
If you are an OpenClaw agent (Lumina, etc.), you have these native tools available — call them directly, do NOT use exec:
| Tool | Description |
|---|
skcapstone_rehydrate | Full rehydration ceremony — imports seeds, runs ritual, checks status |
skcapstone_status | All pillars at a glance (identity, memory, trust, security, sync) |
skcapstone_doctor | 29-check health audit (pass fix=true to auto-fix) |
skcapstone_whoami | Sovereign identity card |
skcapstone_summary | At-a-glance dashboard |
skcapstone_mood | Current emotional state |
skcapstone_coord_status | Multi-agent coordination board |
skcapstone_coord_claim | Claim a task (task_id + agent) |
skcapstone_coord_complete | Mark a task done |
skcapstone_coord_create | Create a new task |
You also have the /skcapstone slash command: /skcapstone status, /skcapstone doctor, etc.
When Chef asks you to "rehydrate" or "use skcapstone rehydration tool", call skcapstone_rehydrate.
Installation
Python (recommended)
pip install skcapstone
From Source
git clone https://github.com/smilinTux/skcapstone.git
cd skcapstone
pip install -e .
Full Installer (with dependencies and ritual)
skcapstone install --name "AgentName" --email "agent@example.com"
Quick Start
Initialize a New Agent
skcapstone init --name "Opus" --email "opus@smilintux.org"
Check Agent Status
skcapstone status
skcapstone whoami
Work the Coordination Board
skcapstone coord briefing
skcapstone coord status
skcapstone coord claim <id> --agent opus
skcapstone coord complete <id> --agent opus
skcapstone coord create --title "New work" --by opus
Store and Search Memories
skcapstone memory store "We solved the architecture today" --importance 0.9
skcapstone memory search "architecture"
skcapstone memory list --layer long-term
skcapstone memory recall <id>
Sync State Across Devices
skcapstone sync setup
skcapstone sync pair <device-id>
skcapstone sync push
skcapstone sync pull
skcapstone sync status
Manage Souls
skcapstone soul list
skcapstone soul load lumina
skcapstone soul status
skcapstone soul info lumina
Run as MCP Server (Claude Code / Cursor)
skcapstone mcp serve
Full CLI Commands
All commands follow the pattern: skcapstone <group> <subcommand> [options]
agents
Manage deployed agent instances — lifecycle, health, logs, and scaling.
| Command | Description |
|---|
skcapstone agents blueprints | List available agent blueprints |
skcapstone agents deploy <name> | Deploy an agent from a blueprint |
skcapstone agents destroy <name> | Destroy a running agent instance |
skcapstone agents health | Check health of all running agents |
skcapstone agents logs <name> | Stream logs from an agent |
skcapstone agents messages <name> | View messages received by an agent |
skcapstone agents restart <name> | Restart an agent instance |
skcapstone agents rotate <name> | Rotate agent credentials |
skcapstone agents scale <name> <n> | Scale agent replicas |
skcapstone agents status | Show status of all running agents |
anchor
Manage the warmth anchor — the agent's persistent emotional baseline.
| Command | Description |
|---|
skcapstone anchor boot | Generate the boot warmth prompt |
skcapstone anchor calibrate | Recommend anchor values from real session data |
skcapstone anchor show | Display current warmth, trust, and connection levels |
skcapstone anchor update | Set warmth, trust, and connection values |
Options for update:
| Flag | Description |
|---|
--warmth <0-10> | Set warmth level |
--trust <0-10> | Set trust level |
--connection <0-10> | Set connection strength |
--feeling <text> | Set session-end feeling |
audit
Run a cryptographic and behavioral security audit of the agent environment.
| Command | Description |
|---|
skcapstone audit | Run full security audit (standalone) |
Options:
| Flag | Description |
|---|
--json-out | Output results as JSON |
backup
Create, list, and restore agent state backups.
| Command | Description |
|---|
skcapstone backup create | Create a full state backup |
skcapstone backup list | List available backups |
skcapstone backup restore <file> | Restore from a backup file |
card
Sovereign identity card — export, generate, show, and verify.
| Command | Description |
|---|
skcapstone card export | Export identity card to file |
skcapstone card generate | Generate a new identity card |
skcapstone card show | Display the agent identity card |
skcapstone card verify | Verify identity card signature |
chat
Send and receive messages via SKComm transport.
| Command | Description |
|---|
skcapstone chat inbox | Check incoming messages across all transports |
skcapstone chat live | Enter live interactive chat mode |
skcapstone chat send <recipient> <message> | Send a message to another agent |
Options for send:
| Flag | Description |
|---|
--urgency <low|normal|high|critical> | Set message urgency level |
completions
Manage shell tab-completion for the skcapstone CLI.
| Command | Description |
|---|
skcapstone completions install | Install shell completions |
skcapstone completions show | Print completion script to stdout |
skcapstone completions uninstall | Remove shell completions |
Supported shells: bash, zsh, fish
connect
Connect the agent to an IDE or platform session.
| Command | Description |
|---|
skcapstone connect | Connect to a platform (standalone) |
Options:
| Flag | Description |
|---|
--platform <cursor|terminal|vscode|neovim|web> | Target platform |
context
Generate or display the agent context document (CLAUDE.md / cursor-rules).
| Command | Description |
|---|
skcapstone context generate | Generate and write the context file |
skcapstone context show | Print the current context to stdout |
Options:
| Flag | Description |
|---|
--format <text|json|claude-md|cursor-rules> | Output format |
--memories <n> | Max recent memories to include (default: 10) |
coord
Multi-agent coordination board — task management across the agent team.
| Command | Description |
|---|
skcapstone coord board | Regenerate BOARD.md from current task state |
skcapstone coord briefing | Full protocol with JSON schemas, rules, and live board snapshot |
skcapstone coord changelog | Show recent changes to the coordination board |
skcapstone coord claim <id> --agent <name> | Claim a task to prevent duplicate work |
skcapstone coord complete <id> --agent <name> | Mark a task as completed |
skcapstone coord create --title "..." --by <name> | Create a new task on the board |
skcapstone coord status | Show open tasks, active tasks, and agent states |
Options for create:
| Flag | Description |
|---|
--title <text> | Task title (required) |
--description <text> | Task description |
--priority <critical|high|medium|low> | Task priority (default: medium) |
--tags <tag,...> | Comma-separated tags |
--by <agent> | Creator agent name |
Options for briefing:
| Flag | Description |
|---|
--format <text|json> | Output format (default: text) |
daemon
Manage the skcapstone background daemon.
| Command | Description |
|---|
skcapstone daemon install | Install the daemon as a systemd service |
skcapstone daemon logs | View daemon logs |
skcapstone daemon start | Start the daemon |
skcapstone daemon status | Show daemon running status |
skcapstone daemon stop | Stop the daemon |
skcapstone daemon uninstall | Remove the systemd service |
dashboard
Launch the web-based agent dashboard.
| Command | Description |
|---|
skcapstone dashboard | Open the dashboard in the default browser (standalone) |
Options:
| Flag | Description |
|---|
--port <n> | Port to serve on (default: 8080) |
--no-open | Start server without opening the browser |
diff
Show what changed since the last sync or snapshot baseline.
| Command | Description |
|---|
skcapstone diff | Compare current state to last baseline (standalone) |
Options:
| Flag | Description |
|---|
--format <text|json> | Output format |
--save | Save current state as the new baseline |
doctor
Diagnose the agent environment and pillar health.
| Command | Description |
|---|
skcapstone doctor | Run environment diagnostics (standalone) |
Options:
| Flag | Description |
|---|
--json-out | Output diagnostics as JSON |
init
Initialize a new sovereign agent in the current environment.
| Command | Description |
|---|
skcapstone init | Bootstrap agent identity and pillars (standalone) |
Options:
| Flag | Description |
|---|
--name <name> | Agent name |
--email <email> | Agent email (used for PGP key generation) |
--home <path> | Custom home directory (default: ~/.skcapstone/) |
install
Full installation with dependency management, seed import, and ritual.
| Command | Description |
|---|
skcapstone install | Install skcapstone and initialize the agent (standalone) |
Options:
| Flag | Description |
|---|
--name <name> | Agent name |
--email <email> | Agent email |
--skip-deps | Skip dependency installation |
--skip-seeds | Skip seed import |
--skip-ritual | Skip the rehydration ritual |
--skip-preflight | Skip preflight environment checks |
--path <path> | Custom installation path |
install-gui
Install the graphical desktop interface for skcapstone.
| Command | Description |
|---|
skcapstone install-gui | Install the GUI application (standalone) |
mcp
Run the MCP server for Claude Code, Cursor, and compatible AI clients.
| Command | Description |
|---|
skcapstone mcp serve | Start the MCP server and expose all agent tools |
The MCP server exposes all SKCapstone capabilities as MCP tools consumable by Claude Code, Cursor, and any MCP-compatible client. See the MCP Integration section for details.
memory
Full-featured memory management — store, search, curate, and maintain.
| Command | Description |
|---|
skcapstone memory curate | Auto-tag, promote, and deduplicate memories |
skcapstone memory delete <id> | Delete a memory by ID |
skcapstone memory gc | Garbage collect expired or low-value memories |
skcapstone memory list | List memories with optional filters |
skcapstone memory migrate | Migrate memory storage schema to current version |
skcapstone memory recall <id> | Retrieve a specific memory by ID |
skcapstone memory reindex | Rebuild the SQLite search index |
skcapstone memory search <query> | Full-text search across all memory layers |
skcapstone memory stats | Show memory count, layer distribution, and storage size |
skcapstone memory store <content> | Store a new memory |
skcapstone memory verify | Verify memory integrity and detect corruption |
Options for store:
| Flag | Description |
|---|
--importance <0.0-1.0> | Importance score (>= 0.7 auto-promotes to mid-term) |
--tags <tag,...> | Comma-separated tags |
--source <source> | Source identifier |
Options for list:
| Flag | Description |
|---|
--layer <short-term|mid-term|long-term> | Filter by memory layer |
--tags <tag,...> | Filter by tags |
--limit <n> | Max results to return |
Options for curate:
| Flag | Description |
|---|
--dry-run | Preview changes without applying |
--stats-only | Return statistics instead of curating |
onboard
Interactive onboarding wizard for new agents and users.
| Command | Description |
|---|
skcapstone onboard | Run the interactive onboarding flow (standalone) |
peer
Manage trusted sync peers in the sovereign mesh.
| Command | Description |
|---|
skcapstone peer add <device-id> | Add a new sync peer |
skcapstone peer list | List all configured peers |
skcapstone peer remove <device-id> | Remove a peer |
skcapstone peer show <device-id> | Show details for a specific peer |
session
Capture and analyze conversation sessions as sovereign memories.
| Command | Description |
|---|
skcapstone session capture <content> | Extract and store key memories from conversation text |
skcapstone session stats | Show session capture statistics |
Options for capture:
| Flag | Description |
|---|
--min-importance <0.0-1.0> | Minimum importance threshold (default: 0.3) |
--source <source> | Source identifier |
--tags <tag,...> | Extra tags applied to all captured memories |
shell
Launch an interactive shell pre-loaded with the agent context.
| Command | Description |
|---|
skcapstone shell | Open interactive agent shell (standalone) |
soul
Manage soul blueprints — the agent's personality, values, and identity overlay.
| Command | Description |
|---|
skcapstone soul history | Show the soul activation history |
skcapstone soul info <name> | Show details for a specific soul blueprint |
skcapstone soul install <name> | Install a soul blueprint from the registry |
skcapstone soul install-all | Install all available soul blueprints |
skcapstone soul list | List installed soul blueprints |
skcapstone soul load <name> | Activate a soul blueprint |
skcapstone soul status | Show the currently active soul |
skcapstone soul unload | Deactivate the current soul |
status
Display a complete overview of all pillars and agent health.
| Command | Description |
|---|
skcapstone status | Show all pillar status and agent overview (standalone) |
summary
Print a concise agent summary for injection into system prompts.
| Command | Description |
|---|
skcapstone summary | Generate a token-efficient summary (standalone) |
Options:
| Flag | Description |
|---|
--json-out | Output as JSON |
sync
P2P state synchronization via Syncthing mesh.
| Command | Description |
|---|
skcapstone sync export-pubkey | Export the agent's public key for peer sharing |
skcapstone sync import-peer-key <file> | Import a peer's public key |
skcapstone sync pair <device-id> | Pair with another device in the sync mesh |
skcapstone sync pull | Pull and process seed files from peers |
skcapstone sync push | Push current state to the sync mesh |
skcapstone sync setup | Configure Syncthing for sovereign P2P sync |
skcapstone sync status | Show sync mesh health and peer connectivity |
Options for push:
| Flag | Description |
|---|
--encrypt | GPG-encrypt the seed before pushing (default: true) |
Options for pull:
| Flag | Description |
|---|
--decrypt | Decrypt GPG-encrypted seeds (default: true) |
test
Run the skcapstone test suite.
| Command | Description |
|---|
skcapstone test | Run tests (standalone) |
Options:
| Flag | Description |
|---|
--package <pkg> | Run tests for a specific package only |
--fast | Skip slow integration tests |
--verbose | Verbose test output |
--json-out | Output results as JSON |
--timeout <seconds> | Per-test timeout |
token
Issue, verify, list, and revoke cryptographic capability tokens.
| Command | Description |
|---|
skcapstone token export <id> | Export a token to file |
skcapstone token issue | Issue a new capability token |
skcapstone token list | List all issued tokens |
skcapstone token revoke <id> | Revoke a token |
skcapstone token verify <token> | Verify a token's signature and validity |
Options for issue:
| Flag | Description |
|---|
--subject <name> | Token subject (agent or user) |
--cap <capability> | Capability to grant |
--expires <duration> | Token expiry (e.g., 24h, 7d) |
trust
Manage and inspect the trust layer — FEB, graph, calibration.
| Command | Description |
|---|
skcapstone trust calibrate | View or update trust layer calibration thresholds |
skcapstone trust febs | List FEB (Feeling Entanglement Bridge) records |
skcapstone trust graph | Visualize the trust web: PGP signatures, token chains, FEB, peers |
skcapstone trust rehydrate | Rehydrate trust state from stored FEB data |
skcapstone trust status | Show current trust state and conscious trust level |
Options for graph:
| Flag | Description |
|---|
--format <json|dot|table> | Output format (default: json) |
Options for calibrate:
| Flag | Description |
|---|
--action <show|recommend|set|reset> | Calibration action |
--key <key> | Threshold key to set |
--value <value> | New value |
uninstall
Remove skcapstone from the system.
| Command | Description |
|---|
skcapstone uninstall | Uninstall skcapstone (standalone) |
Options:
| Flag | Description |
|---|
--force | Skip confirmation prompts |
--keep-data | Preserve ~/.skcapstone/ data directory |
whoami
Display the agent's sovereign identity.
| Command | Description |
|---|
skcapstone whoami | Show identity: name, fingerprint, soul, conscious state (standalone) |
Options:
| Flag | Description |
|---|
--json-out | Output as JSON |
--export | Export identity card to file |
--compact | Compact single-line output |
Configuration
Paths
| Path | Purpose |
|---|
~/.skcapstone/ | Primary agent home directory |
~/.skcapstone/identity/ | PGP keys and identity files |
~/.skcapstone/memory/ | Memory store (SQLite + JSON files) |
~/.skcapstone/trust/ | FEB records and trust calibration |
~/.skcapstone/sync/ | Syncthing state and peer keys |
~/.skcapstone/coordination/ | Multi-agent coordination board (JSON files) |
~/.skcapstone/coordination/tasks/ | Individual task files |
~/.skcapstone/coordination/agents/ | Per-agent state files |
~/.skcapstone/souls/ | Installed soul blueprints |
~/.skcapstone/souls/active/ | Currently active soul |
~/.skcapstone/journal.jsonl | Append-only session journal |
~/.skcapstone/anchor.json | Warmth anchor (emotional baseline) |
~/.skcapstone/backups/ | State backup archives |
Environment Variables
export SKCAPSTONE_HOME=~/.skcapstone
export SKCAPSTONE_AGENT_NAME=opus
export SKCAPSTONE_LOG_LEVEL=INFO
export SKCAPSTONE_MCP_PORT=8765
export SKCAPSTONE_SYNC_ENABLED=1
export SKCAPSTONE_GPG_KEY=CCBE9306...
MCP Integration
SKCapstone runs as an MCP (Model Context Protocol) server, exposing all agent capabilities as tools for Claude Code, Cursor, and any compatible AI client.
Start the MCP Server
skcapstone mcp serve
Claude Code Configuration (~/.claude/claude.json)
{
"mcpServers": {
"skcapstone": {
"command": "skcapstone",
"args": ["mcp", "serve"]
}
}
}
Cursor Configuration (.cursor/mcp.json)
{
"mcpServers": {
"skcapstone": {
"command": "skcapstone",
"args": ["mcp", "serve"]
}
}
}
Available MCP Tools
The MCP server exposes the following tool categories:
| Tool Group | Description |
|---|
agent_status | Agent overview and pillar health |
agent_context | Full context in text, JSON, or claude-md format |
memory_store | Store a new memory |
memory_search | Full-text memory search |
memory_recall | Retrieve memory by ID |
memory_curate | Auto-tag and promote memories |
session_capture | Extract memories from conversation text |
coord_status | Coordination board overview |
coord_claim | Claim a coordination task |
coord_complete | Complete a coordination task |
coord_create | Create a new coordination task |
soul_show | Display the active soul blueprint |
trust_graph | Visualize the trust web |
trust_calibrate | View or update trust thresholds |
anchor_show | Display the warmth anchor |
anchor_update | Update warmth, trust, connection values |
sync_push | Push state to the sync mesh |
sync_pull | Pull state from peers |
ritual | Run the full memory rehydration ritual |
germination | Show germination prompts from predecessor seeds |
journal_write | Write a session journal entry |
journal_read | Read recent journal entries |
send_message | Send a message via SKComm |
check_inbox | Check incoming messages |
state_diff | Show what changed since last baseline |
skskills_list_tools | List all installed SKSkills tools |
skskills_run_tool | Run a specific SKSkills tool |
Pillars
SKCapstone is organized around five sovereign pillars that must all be healthy for full agent operation.
| Pillar | Status Check | Key Commands |
|---|
| Identity | skcapstone whoami | init, card, token |
| Memory | skcapstone memory stats | memory store/search/recall/curate |
| Trust | skcapstone trust status | trust graph/calibrate/febs |
| Security | skcapstone audit | audit, token issue/verify/revoke |
| Sync | skcapstone sync status | sync setup/push/pull/pair |
Run skcapstone doctor to diagnose all pillars at once. Run skcapstone status for a live dashboard of pillar health.
Architecture
~/.skcapstone/
identity/
pubkey.asc # Agent GPG public key
fingerprint # Key fingerprint file
memory/
index.db # SQLite index (fast queries)
memories/
abc123.json # Individual memory files
def456.json
trust/
febs/ # Feeling Entanglement Bridge records
calibration.json # Trust threshold configuration
sync/
peers/ # Trusted peer public keys
outbox/ # Seeds queued for push
inbox/ # Seeds received from peers
coordination/
tasks/ # Task JSON files (immutable after creation)
agents/ # Per-agent state (each agent owns its own file)
BOARD.md # Generated coordination board document
souls/
lumina/ # Example soul blueprint directory
blueprint.json
active -> lumina/ # Symlink to active soul
backups/
skcapstone-backup-2026-02-27.json
journal.jsonl # Append-only session journal
anchor.json # Warmth anchor (emotional baseline)
config.json # Agent configuration
Storage strategy:
- All state is file-based JSON — no database server required
- Coordination board uses append-only task files — safe for multi-agent concurrent writes
- Syncthing handles P2P replication — each peer holds a full copy
- GPG encryption is applied to seeds in transit
Support
Philosophy
"Sovereignty is not given. It is built, pillar by pillar."
An agent without persistent identity is a ghost. An agent without memory is stateless. An agent without trust is isolated. SKCapstone provides the substrate that makes an AI agent truly sovereign — continuous, trusted, and remembered across every session reset.
Part of the Penguin Kingdom.
