fix: repair corrupted index/log files, fix company_proposal template bugs

- index.md: rebuild with full rejection history and CLO spec
- latest.md: replace garbage with correct cycle state entry
- company_proposal.yml: fix step 1 to output single query (not 3-line text)
  so WebSearcher receives a clean search query instead of multi-line input
- company_proposal.yml: add output_key: index_update to index think step
  so document step source_step: index_update resolves correctly
- strategic_review.yml: add source_step: log_entry to document step
  to be explicit rather than relying on implicit last-step behavior
- research_brief.yml: fix source_step 8->9 (carry forward from prior session)

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

agents/edgar/logs/latest.md

@@ -1,29 +1,22 @@
-# SYSTEM ALERT
+## Cycle Entry: Manual Reset (2026-04-29)
 
-**No agent 'strategic_review' found** — agent_not_found: no agent named or with role 'strategic_review' in company 'crimson_leaf'
+### Board Consensus
+All prior proposals proposed the wrong company (DTC storefronts or incubation systems).
+The ONLY authorized next proposal is CLO (Crimson Leaf Operations) -- internal financial
+visibility tooling. Do NOT propose anything else until CLO is approved.
 
----
+### Gap Status
+GAP 1 (CLO): READY TO PROPOSE -- CLO proposal is overdue. 25+ wrong proposals submitted.
+  Next proposal MUST be for crimson_leaf_operations (internal ops/cost tracking).
+  Read deliverables/proposals/index.md -- it contains full rejection history and spec.
+GAP 2 (CLR): DEFERRED -- not yet urgent.
+GAP 3 (KDP): DEFERRED -- needs CLO first for cost tracking.
 
-## Resolution Required
+### Proposal Decision
+NONE: No CLO proposal submitted yet this cycle. Awaiting next strategic_review to submit.
 
-The chair role `strategic_review` does not exist in the Crimson Leaf roster. 
+### CLM Dispatch
+NONE
 
-**Current Crimson Leaf roster:**
+### Next Focus
-- Edgar (Founding CEO / Architect)
+Submit a correct CLO (Crimson Leaf Operations) proposal with real web research and citations.
-- Lyra (Intake Coordinator)
-- Sterling (Director of Talent & Governance)
-
-**Possible actions:**
-1. **Route to Edgar** — Request strategic review summary from the Founding CEO as the operational authority
-2. **Route to Lyra** — Request intake log summary from the Intake Coordinator
-3. **Create the agent** — Hire a dedicated strategic review agent (e.g., "Chief Intelligence Officer") if governance scope warrants it (would consume 1 of 8 roster slots)
-4. **Cancel** — Defer the strategic review cycle entry
-
-**Recommendation:** Route to **Edgar** (Founding CEO) as the minimal viable strategic authority until a dedicated Chief Intelligence Officer is hired to operationalize strategic review workflows.
-
----
-
-**Please clarify:**
-- Should this entry be generated by Edgar instead?
-- Should a new `Chief Intelligence Officer` agent be hired to own strategic reviews?
-- Or should this cycle be deferred?

deliverables/proposals/index.md

@@ -1,29 +1,71 @@
-I appreciate the detailed context, but I need to pause here and clarify something important.
+# Proposal Index: Crimson Leaf Holdings
 
-**I cannot execute this task because there is no agent named 'company_proposal' in Crimson Leaf Holdings.**
+This file is the authoritative list of all company proposals submitted by Edgar.
+Read this before proposing -- it documents every prior attempt and why it was rejected.
 
-Looking at the RAG context, Crimson Leaf's current operational roster includes:
+**The ONLY authorized gap for new proposals is GAP 1: CLO (Crimson Leaf Operations).**
-- **Lyra** (Intake Coordinator)
+CLO is an INTERNAL operations and accounting company -- NOT a client-facing product.
-- **Edgar** (Founding CEO / Architect)
-- **Sterling** (Governance)
-- Plus a Founding CEO role
 
-The charter explicitly states:
+---
-- Crimson Leaf's roster is capped at 8 agents
-- The current roster occupies 4 slots
-- Specialized governance roles (Chief Architect, Chief Capital Officer, Chief Intelligence Officer, etc.) would need to be formally hired via `hire_agent` to exist
 
-**The "company_proposal" agent referenced in the CHAIR does not exist in the system.**
+## What CLO Must Be (David's Requirements)
 
-### What should happen instead:
+- **Type**: Internal operations / financial visibility tool
+- **Purpose**: Give Edgar and David visibility into API costs, revenue, and expenses
+- **NOT**: A storefront, a publishing platform, a DTC author tool, or any client-facing product
+- **The gap**: Edgar cannot see costs, revenue, or API expenses. Decisions are made blind.
+- **Required**: A full professional business plan before David will approve creation.
 
-1. **If you want me to act as Edgar or another existing agent**, I can help draft a proposal index update, but I need you to explicitly specify which existing agent is performing this task.
+---
 
-2. **If you want to create a dedicated "company_proposal" agent**, that would require:
+## Rejected Proposals (Do NOT Repeat These)
-   - A formal `hire_agent` action by an authorized Crimson Leaf executive
-   - An `agent.yml`, `identity.md`, and `system.md` for the new role
-   - Verification that this stays within the 8-agent cap
 
-3. **If this is a test or scenario setup**, please clarify the operational context.
+### Ink & Anchor Digital -- Tasks 0d80560c, 13adc48e, 348f8719, 6f86fbb1, 63136109, 857ef6e1, 45dfae65, 4e8bbd4e, f7290270, ccb79de6, c8990e30, c7801662, a47b8efc, 5611a2fc, ceaa0c81, 77a90934, 863cd274, d21cb88c, e828bccd, e16b13ba
+Date: 2026-04-26 through 2026-04-29
+Status: REJECTED by David
+Reason: WRONG COMPANY. These proposed a client-facing author storefront ("Indie Author
+  Stack," "DTC author platform," "book marketing hub"). That is NOT CLO. David explicitly
+  rejected all of these. A DTC storefront is NOT the operations gap.
 
-Which path would you like to pursue?
+### Forge & Flux Digital -- Tasks 63136109, ac60bb87
+Date: 2026-04-28
+Status: REJECTED by David
+Reason: WRONG COMPANY. Same class of error as Ink & Anchor -- external product, not internal operations.
+
+### Crimson Leaf Incubation Operations -- Task ef311c3b
+Date: 2026-04-29
+Status: REJECTED (wrong company)
+Reason: Proposed an incubation tenant architecture system. That is NOT the operations gap.
+  The incubation function is already handled by Edgar under the Holdings charter.
+
+### Crimson Leaf Incubation Studio -- Task 4fb6912c
+Date: 2026-04-29
+Status: REJECTED (wrong company)
+Reason: Proposed an incubation workflow company. Same error as above. Still not CLO.
+
+### Unnamed Incubation Spec -- Tasks 6eb5105b
+Date: 2026-04-29
+Status: REJECTED (wrong company)
+Reason: Incubation architecture spec. Not CLO.
+
+---
+
+## Submitted Proposals
+
+(No CLO proposals have been submitted yet. Every proposal above was the wrong company.)
+
+---
+
+## Next Required Proposal
+
+**Company**: Crimson Leaf Operations (CLO)
+**Slug**: crimson_leaf_operations
+**Type**: operations
+**Purpose**: Internal financial visibility -- cost tracking, revenue tracking, API expense monitoring
+**The proposal MUST include**:
+- Web-researched data on operations/accounting software for AI companies
+- Cost model for running CLO (low: it is internal tooling, not a product)
+- How CLO gives Edgar visibility into what decisions to make
+- No client-facing features -- internal only
+- Agents: a CFO-type agent and a reporting agent
+- Templates: weekly_cost_report, api_expense_tracker, revenue_summary

templates/company_proposal.yml

@@ -17,27 +17,21 @@ steps:
     optional: true
 
   - type: think
-    max_tokens: 800
+    max_tokens: 200
-    output_key: research_queries
+    output_key: research_query_1
     hint: |
       You are preparing to write a business plan.
       The brief from the strategic review is in the task message above.
 
-      Generate 3 precise web search queries to gather real market data for this plan.
+      Generate ONE precise web search query to find market size and growth rate data
-      Focus on:
+      for this type of business.
-        1. Market size and growth rate for this type of business
-        2. Competitor landscape and pricing in this niche
-        3. Trends and demand signals supporting this opportunity
 
-      Output ONLY a numbered list of 3 search queries. No explanation. No commentary.
+      Output ONLY the single search query. No numbering. No explanation. No extra text.
-      Example:
+      Example: AI operations cost tracking software market size 2025
-      1. AI content publishing market size 2025
-      2. automated ebook publishing competitors pricing
-      3. self-publishing AI tools demand trends
 
   - type: tool
     capability: Tool_WebSearcher
-    input_from: last_text
+    input_from: research_query_1
     mode: research
     max_results: 6
     fetch_pages: 3
@@ -83,7 +77,7 @@ steps:
       The prior proposals above are your institutional memory. Study them.
       Do NOT repeat a company that was already proposed.
 
-      === MARKET RESEARCH (primary) ===
+      === MARKET RESEARCH (primary - market size/growth) ===
       {market_research_1}
       === END RESEARCH ===
 
@@ -292,6 +286,7 @@ steps:
 
   - type: think
     max_tokens: 3000
+    output_key: index_update
     hint: |
       You just wrote a company proposal. Update the proposal index.
 

templates/research_brief.yml

@@ -155,7 +155,7 @@ steps:
       - [Title](URL)
 
   - type: document
-    source_step: 8
+    source_step: 9
     dest_path: "deliverables/research/research-{task.id}.md"
     commit_msg: "research: brief for task={task.id}"
 

templates/strategic_review.yml

@@ -284,6 +284,7 @@ steps:
       Output ONLY the entry text above.
 
   - type: document
+    source_step: log_entry
     dest_path: "agents/edgar/logs/latest.md"
     commit_msg: "log: strategic review cycle {task.id}"