fix: add proposal index, wire prior proposals into edgar review + proposal templates, unblock CLO proposal

agents/edgar/strategic_log.md

@@ -22,12 +22,12 @@ Per-cycle entries are written to agents/edgar/logs/latest.md.
 GAP 1 (CLO - Crimson Leaf Operations): AUTHORIZED BY DAVID -- highest priority.
   Edgar cannot see costs, revenue, or API expenses.
   A full business plan is required before David will create the company.
-  Status: PROPOSAL SUBMITTED -- AWAITING DAVID'S APPROVAL.
+  Status: READY TO PROPOSE -- but do it correctly this time.
-  NOTE: 20+ proposals for "Ink & Anchor Digital / Forge & Flux Digital" have been
+  NOTE: 20 prior proposals were all rejected. They proposed a client-facing author DTC
-  written to deliverables/proposals/ between April 26-28, 2026. DO NOT submit another
+  storefront ("Ink & Anchor Digital"). That is the WRONG company. David rejected them all.
-  proposal. The pipeline kept re-proposing due to a log memory bug (now fixed).
+  The next proposal MUST be for CLO as an INTERNAL operations/accounting company.
-  Edgar must WAIT for David to approve or reject before any new proposal is generated.
+  Read deliverables/proposals/index.md before proposing -- it details what was wrong
-  RULE: If this log says "AWAITING DAVID'S APPROVAL", output PROPOSAL_NEEDED: none.
+  with every prior attempt and what the correct proposal must contain.
 
 GAP 2 (CLR - Crimson Leaf Research): DEFERRED -- not yet urgent.
 GAP 3 (KDP automated pipeline): DEFERRED -- needs CLO first for cost tracking.

deliverables/proposals/index.md

@@ -0,0 +1,62 @@
+# Proposal Index: Crimson Leaf Holdings
+
+This file is READ at the start of each company_proposal run and UPDATED at the end.
+It is the authoritative list of every proposal ever submitted.
+Edgar must read this before writing any new proposal.
+
+---
+
+## Submitted Proposals
+
+### BATCH 1 (April 26-28, 2026): "Ink & Anchor Digital" variants -- ALL REJECTED BY DAVID
+These 20 proposals were generated by a template bug that ignored the mission charter.
+They all proposed a client-facing indie author DTC storefront called "Ink & Anchor Digital"
+or "Forge & Flux Digital." This is NOT what the mission charter authorizes.
+
+David's verdict: "None of the proposals are worth reading. I am looking for full business plans."
+Status: REJECTED. Do NOT repropose this concept in any form.
+
+Task IDs (all rejected, do not reference):
+- 0d80560c, 13adc48e, 348f8719, 45dfae65, 4e8bbd4e, 63136109, 6f86fbb1, 857ef6e1
+- ccb79de6, c8990e30, c7801662, a47b8efc, f7290270, 5611a2fc, d21cb28c, ac60bb87
+- 77a90934, ceaa0c81, 1550e9e3, 863cd274
+
+Reason for rejection: Proposed the wrong company. Mission charter Gap 1 is CLO --
+Crimson Leaf Operations -- an INTERNAL financial visibility company, not a
+client-facing author services agency.
+
+---
+
+## What the Next Proposal Must Be
+
+GAP 1 (AUTHORIZED BY DAVID, HIGHEST PRIORITY):
+  Company: Crimson Leaf Operations (CLO)
+  Slug: crimson_leaf_operations
+  Mission: Give Edgar financial visibility. Track API costs, revenue, and P&L
+           across all subsidiaries. Without CLO, Edgar is flying blind.
+  This is an INTERNAL operations company. It does NOT serve clients.
+  It does NOT build websites. It does NOT serve indie authors.
+  It tracks money: what is spent, what is earned, what the margins are.
+
+The next proposal MUST be for CLO. Build a serious, detailed plan:
+- How CLO tracks costs per company, per task, per model
+- How CLO generates weekly P&L reports for David
+- What agents run CLO (an internal accountant/analyst)
+- What templates it uses (cost_report, revenue_summary, etc.)
+- How it integrates with the PAE task system to capture cost data
+- 90-day success criteria: weekly P&L delivered, cost-per-chapter tracked, etc.
+
+Do NOT pitch a different company. Do NOT rename it. Do NOT make it client-facing.
+
+---
+
+## Instructions for the company_proposal template
+
+When you finish writing a proposal, append a new entry to this file under
+"## Submitted Proposals" with:
+- The company name and slug
+- The task ID
+- A 2-3 sentence summary of what was proposed
+- David's verdict (if known -- leave blank if awaiting)
+
+This keeps Edgar's institutional memory intact across cycles.

templates/company_proposal.yml

@@ -9,6 +9,13 @@ sections:
   - message
   - instructions
 steps:
+  - type: tool
+    action: git_read_file
+    params:
+      path: "deliverables/proposals/index.md"
+    output_key: prior_proposals
+    optional: true
+
   - type: think
     max_tokens: 2000
     output_key: exec_summary
@@ -20,6 +27,14 @@ steps:
       the brief. Do NOT rename, rebrand, or creatively reinterpret the company.
       The brief was written from the mission charter -- it is authoritative.
 
+      === PRIOR PROPOSALS (read this before writing) ===
+      {prior_proposals}
+      === END PRIOR PROPOSALS ===
+
+      The prior proposals above are your institutional memory. Study them.
+      Do NOT repeat a company that was already proposed. Do NOT reuse names,
+      slugs, or concepts that were rejected. Build something new and better.
+
       Write the EXECUTIVE SUMMARY section. Include:
 
       1. PROPOSED COMPANY
@@ -206,6 +221,32 @@ steps:
     dest_path: "deliverables/proposals/proposal-{task.id}.md"
     commit_msg: "proposal: company_proposal task={task.id}"
 
+  - type: think
+    max_tokens: 600
+    output_key: index_update
+    hint: |
+      You just wrote a company proposal. Update the proposal index.
+
+      Read the current index content:
+      {prior_proposals}
+
+      Append a new entry at the bottom of the "## Submitted Proposals" section.
+      Use this exact format:
+
+      ### [Company Full Name] -- Task {task.id}
+      Date: 2026-04-29
+      Status: AWAITING DAVID'S APPROVAL
+      Summary: [2-3 sentences: what was proposed, what gap it fills, key differentiator
+                from any prior proposals]
+
+      Output the COMPLETE updated index file content from the very beginning.
+      Do not truncate. Do not skip sections. Include all prior entries plus the new one.
+
+  - type: document
+    source_step: index_update
+    dest_path: "deliverables/proposals/index.md"
+    commit_msg: "index: add proposal {task.id} to proposal index"
+
   - type: reply
     target: channel
     channel_name: "crimson_leaf:general"

templates/strategic_review.yml

@@ -41,6 +41,13 @@ steps:
     output_key: latest_log_entry
     optional: true
 
+  - type: tool
+    action: git_read_file
+    params:
+      path: "deliverables/proposals/index.md"
+    output_key: prior_proposals_index
+    optional: true
+
   - type: think
     max_tokens: 8000
     rotate_participants: true
@@ -60,6 +67,9 @@ steps:
 
       === LAST CYCLE ENTRY ===
       {latest_log_entry}
+
+      === PRIOR PROPOSALS (all proposals ever submitted) ===
+      {prior_proposals_index}
       ===
 
       Analyze from YOUR perspective:
@@ -162,6 +172,11 @@ steps:
       - 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_index}
+      ===
 
       Output ONLY the brief text. No labels. No preamble.