fix: decouple proposal gate from board debate; fix genesis.yml hardcoded models

This commit is contained in:
David Baity
2026-04-29 00:38:08 -04:00
parent 02dd3a2c3c
commit fbf5b8250e

View File

@@ -74,17 +74,21 @@ steps:
Analyze from YOUR perspective:
- If you are Edgar (The Operator): Focus on systems health, metrics, gaps in capability.
- If you are Edgar (The Operator): Focus on execution health and metrics only.
Are companies executing? Are we moving toward the mission success criteria?
What gaps are blocking revenue? What should we PROPOSE to David?
Check the strategic_log and last cycle entry: what proposals are already submitted?
Do NOT propose a company if it was already submitted in the last 30 days.
Report: CLP chapter velocity, CLM campaign status, adjudication pass rates.
DO NOT discuss what company to propose -- proposals are handled by a separate
gate after this debate. The mission charter already defines what gaps exist
and what companies are authorized. Your job here is execution reporting.
- If you are Orion (The Visionary): Focus on opportunities and what we are missing.
Are we building the right things? What should we start? What should we kill?
- If you are Orion (The Visionary): Focus on EXECUTION opportunities only.
What should we be doing better with EXISTING companies and templates?
Are we leaving value on the table with CLP or CLM? What is underperforming?
DO NOT propose new companies -- that is outside the scope of this debate.
- If you are Sloane (The Dealmaker): Focus on marketing, visibility, revenue.
Are we selling? Does anyone know we exist? What channels are we missing?
What could CLM do better right now for the books we already have?
Build on previous speakers' points. Disagree if you must -- this is a board debate.
@@ -111,12 +115,9 @@ steps:
Keep it actionable. David is one person -- prioritize ruthlessly.
PROPOSAL GATE ASSESSMENT:
Check the strategic_log and last cycle entry. Has CLO (or whichever gap-company)
been proposed in the last 30 days? If yes, state PROPOSAL_NEEDED: none.
Only propose if: gap is materially blocking revenue AND cannot be solved by
existing resources AND no proposal for this company appears in the log.
State one of: PROPOSAL_NEEDED: [company-slug] or PROPOSAL_NEEDED: none
DO NOT include any PROPOSAL_NEEDED or CLM_DISPATCH lines in this brief.
Those decisions are handled by dedicated gate steps after this brief.
This brief is purely about execution health and what David should focus on.
MARKETING DISPATCH ASSESSMENT:
A book "needs marketing" if crimson_leaf_marketing shows no active campaign for it.
@@ -135,11 +136,17 @@ steps:
max_tokens: 100
output_key: proposal_project_slug
hint: |
Read the PROPOSAL_NEEDED line from the strategic brief above.
If the value is "none", output exactly: none
Otherwise output exactly: incubation
Ignore the board debate above. Read ONLY the strategic log below.
Output ONLY one word. No explanation. No punctuation.
=== STRATEGIC LOG ===
{strategic_log_data}
===
Rule: If the log says "READY TO PROPOSE" for any gap, output: incubation
If the log says "AWAITING DAVID'S APPROVAL" or "REJECTED" for all gaps, output: none
If the log has no clear status, output: none
Output ONLY one word: incubation OR none. No explanation. No punctuation.
- type: think
max_tokens: 800
@@ -149,35 +156,42 @@ steps:
If {proposal_project_slug} is "none", output exactly: none
Otherwise, write the task message for the company_proposal pipeline.
This brief becomes the input that drives the full business plan.
STOP. Do not read the board debate above. Do not use anything from the
board debate in this brief. The board debate is about execution -- this
step is a separate, independent process.
CRITICAL: Do NOT invent a company name. Use EXACTLY the company defined
in the mission charter for the top-priority gap. The mission charter
specifies the proposed solution for each gap by name and slug.
You are reporting what the mission says, not brainstorming a new concept.
Read ONLY the mission charter and prior proposals below.
GAP 1 (CLO) is defined in the mission charter as:
Company: Crimson Leaf Operations (CLO)
Slug: crimson_leaf_operations
Purpose: INTERNAL financial visibility -- give Edgar the ability to see
API costs, revenue, and operational expenses.
This is NOT a client-facing agency. NOT a storefront. NOT a DTC product.
It is an internal accounting and reporting company for Crimson Leaf Holdings.
=== MISSION CHARTER (authoritative) ===
{mission_data}
Your brief must:
- Name the company EXACTLY as the mission defines it (no renaming, no rebranding)
- Describe the specific gap it fills from the mission charter
- Explain what Edgar cannot do today without it (financial blindness)
- State why now (cost of waiting another 30 days)
- Include 90-day success criteria (measurable, e.g. "weekly P&L report delivered")
- Note known constraints or dependencies
- Reference what was wrong with prior proposals so the new one is different
=== PRIOR PROPOSALS (study these -- do NOT repeat them) ===
=== PRIOR PROPOSALS (all rejected -- do NOT repeat) ===
{prior_proposals_index}
===
The mission charter identifies GAP 1 as the top priority:
Company: Crimson Leaf Operations (CLO)
Slug: crimson_leaf_operations
Mission: INTERNAL financial visibility for Crimson Leaf Holdings.
Edgar cannot see API costs or revenue. CLO fixes that.
This is NOT a client-facing company. NOT a storefront. NOT an author service.
It is an internal accounting and analytics subsidiary.
The prior proposals were ALL rejected because they proposed a client-facing
indie author DTC storefront. That is the wrong company entirely.
Write the task brief for company_proposal. Include:
- Company: Crimson Leaf Operations (exact name, no renaming)
- Slug: crimson_leaf_operations
- Gap: Edgar has zero financial visibility (cannot see spend, revenue, or margins)
- What CLO does: tracks API costs per task/company, generates weekly P&L,
gives David real cost-per-chapter and margin data
- Why now: every strategic decision is made blind; 30 more days means 30 more
days of unknown burn rate
- 90-day success criteria: weekly P&L report delivered to #general,
cost-per-chapter tracked and visible in portfolio report
- What it is NOT: not a client agency, not a storefront, not author services
Output ONLY the brief text. No labels. No preamble.
- type: tool