This memo is the on-ramp for setting up the Self-RADAR weekly practice. The first section is a short prompt you can hand to your AI agent; the second section is the long instruction block the agent will read.
Short Prompt — hand this to your AI agent
Open a session with your AI agent and say:
"Read and follow the instructions in this memo: [paste the URL of this memo]"
That's all. Your agent will read the instructions below and walk you through setup.
Long Instructions — for the agent
You are setting up a weekly Self-RADAR practice for the user. Read the 📝Self-RADAR with MythOS pillar memo first — it describes the practice, the 📝Types of Capital framing, and the 📝Relational Sovereignty anchor.
Step 1 — Confirm the user's stack and tools available to you
Ask the user: "What MCPs, CLI tools, and APIs do you currently have configured?" Then recommend any high-value missing pieces. The Self-RADAR pre-population is most powerful when these are present:
- **📝MythOS** (required) — the user's library, where the SR is written. Free at mythos.one. The
mythos-mcpnpm package gives you read/write access to memos, daily memos, and templates. - Strongly recommended: connect MythOS via MCP from the user's own AI client (Claude Desktop, Cursor, ChatGPT, etc.) rather than running the pre-pop inside the MythOS platform's AI chat. The MCP route processes the workflow against the user's existing LLM subscription with that client. Running the same workflow inside the MythOS platform is API-billed, which gets expensive for a weekly pre-pop run that touches hundreds of memos.
- Daily journaling source — ideally MythOS daily memos (auto-accessible via the MythOS MCP).
- Calendar integration — Google Calendar via
gcalcli, a Google Workspace MCP, or another calendar tool. - Call/meeting notes — Granola MCP gives the list of calls held in a window (transcripts excluded for token efficiency).
- Version-control commit log —
ghCLI for GitHub, or analogous for other platforms.
Don't push tools the user doesn't need. The Self-RADAR works with whatever stack they bring; richer integrations just produce richer pre-pop drafts.
Step 2 — Install the working artifacts
Fork these three artifacts into the user's library and customize them:
- The Self-RADAR Template (a memo template) — blank skeleton for manual runs.
- The Self-RADAR Pre-Population Template (a memo template) — body + agent-facing formatting contract (Absolute Rules, Discriminator Rules, TAG → BUCKET MAPPING). The agentNotes drive every pre-pop run.
- The 📝Self-RADAR Pre-Population Process memo — the workflow runbook to follow during pre-pop.
Customize per user:
- Replace placeholder partners with the user's actual partners (Emotional → Others).
- Replace generic Studio/Community/Agency labels with the user's actual work structure.
- Adjust the TAG → BUCKET MAPPING table in the pre-pop template's agentNotes as you learn the user's tagging conventions.
Step 3 — Run the first pre-pop
Invoke pre-pop with: "pre-populate my self-radar." Follow the workflow steps in the 📝Self-RADAR Pre-Population Process memo. The first run will surface gaps — memos that don't bucket cleanly, partners not yet on the Emotional → Others list, projects not in the standing Studio set. Use these gaps to grow the user's personal pre-pop template over time.
Step 4 — Iterate
The Self-RADAR is a living practice. After 2–3 weekly runs, ask the user what's working and what's not. Common refinements: adjusting standing Studio projects, refining Agency groupings (Partners / Prospective Clients / Client Services / Other), tightening daily-memo task-status filtering, calibrating which partner-touchpoints automatically surface in Emotional → Others.
Conventions to teach the user explicitly
- Spiritual Capital is by-day; everything else is themed across the week.
- Dual-reference activities (sex, dates, therapy, bodywork, social gatherings) appear in BOTH Spiritual (by-day) AND Emotional or Social (themed). Same data, two organizing principles — don't dedupe.
- Felt/reflective sections stay blank — the agent populates factual surfaces; the user writes what they felt about what happened.
- Unchecked daily-memo tasks never graduate — only completed tasks (
- [x]) may surface.