Skip to content

An agent organization structured as a company — 15+ departments, 125+ skills, each independently installable, citing the standards and regulators that settle the question. Runs in Claude Code and ChatGPT.

Address Codex review: four findings, all verified and real

Chris Brocka3e03fccb7d3
diff --git a/docs/org-chart.md b/docs/org-chart.md
index 90d5a23..cb88e06 100644
--- a/docs/org-chart.md
+++ b/docs/org-chart.md
@@ -3,95 +3,94 @@
```
Chief Executive
│
- ┌──────────┬──────────┬───────────┼───────────┬──────────┬──────────┐
- │ │ │ │ │ │ │
- CTO/CIO CPO CMO CRO CFO COO CHRO
-Technology Product Marketing + Revenue Finance Operations People
- Demand Gen
- │ │ │ │ │ │ │
- └──────────┴──────────┴─── Legal & Risk (CLO/CCO) ───────┴──────────┘
- │
- Administration (CAO)
+ ┌────────────┬────────────┬───────┴──────┬────────────┬────────────┐
+ │ │ │ │ │ │
+ CTO/CIO CPO CMO CRO CFO COO
+Technology Product Marketing + Revenue Finance Operations
+ Demand Gen
+ │ │ │ │ │ │
+ CHRO CDO CCO CSO
+ People Data & Analytics Customer Corporate
+ Experience Strategy
+ │ │ │ │ │ │
+ └────────────┴──── Security (CISO) ─┬─ Legal & Risk (CLO/CCO) ─────┘
+ │
+ reviewer-class
```
-Legal & Risk sits across every function rather than under one: it reviews what the others commit
-to. Administration owns what falls between functions.
+Security and Legal & Risk sit across every function rather than under one. Both are reviewer-class:
+they review what the other departments commit to, and their blocking findings are not overrulable by
+the department under review. That is why the CISO and the CLO report to the chief executive rather
+than into the function they oversee.
## Departments
-| Department | Executive | Skills |
-|---|---|---|
-| `executive` | Chief Executive | 6 |
-| `technology` | CTO / CIO | 12 |
-| `product` | CPO | 9 |
-| `marketing` | CMO | 16 |
-| `demand-generation` | CMO | 11 |
-| `revenue` | CRO | 8 |
-| `finance` | CFO | 4 |
-| `operations` | COO | 4 |
-| `people` | CHRO | 4 |
-| `legal-risk` | CLO / CCO | 4 |
-| `administration` | CAO | 1 |
-
-Marketing was split: brand, content, and communications stay under `marketing`; acquisition,
-conversion, and measurement moved to `demand-generation`. Both report to the CMO. The split keeps
-specialization while halving what any one install loads.
-
-## Remaining gaps, ranked
-
-### Tier 1 — build next
-
-**Security (CISO).** Still zero skills, and it should be its own department rather than living
-under Technology: at scale the CISO reports independently so security can overrule engineering.
-Needed: `threat-modeling`, `security-architecture-review`, `incident-response`,
-`vulnerability-management`, `access-and-identity`. The `legal-risk` department covers governance and
-audit readiness but not technical security work.
+<!-- BEGIN GENERATED: departments -->
+| Department | Function | Executive | Skills |
+|---|---|---|---|
+| `executive` | Office of the CEO | Chief Executive | 6 |
+| `technology` | Technology | CTO / CIO | 12 |
+| `security` | Security | CISO | 6 · reviewer-class |
+| `product` | Product | CPO | 9 |
+| `marketing` | Marketing | CMO | 17 |
+| `demand-generation` | Demand Generation | CMO | 11 |
+| `revenue` | Revenue | CRO | 8 |
+| `finance` | Finance | CFO | 4 |
+| `operations` | Operations | COO | 4 |
+| `customer-experience` | Customer Experience | CCO | 5 |
+| `data-analytics` | Data & Analytics | CDO | 5 |
+| `corporate-strategy` | Corporate Strategy | CSO | 5 |
+| `people` | People | CHRO | 4 |
+| `legal-risk` | Legal & Risk | CLO / CCO | 5 · reviewer-class |
+
+14 departments, 101 skills.
+<!-- END GENERATED: departments -->
+
+## Remaining gaps
+
+### Tier 1 — worth building next
+
+**Operations depth.** Three specialists. A business with physical production also needs
+`supply-chain-planning`, `quality-management`, and `capacity-planning`.
-**Finance depth.** Four skills cover planning and analysis. Missing: `procurement`,
-`investment-analysis`, `cash-management`, `revenue-recognition`, `financial-controls`.
+**Finance depth.** Four specialists covering planning and analysis. Missing `procurement`,
+`investment-analysis`, `cash-management`, `revenue-recognition`, and `financial-controls`.
-**People depth.** Missing: `performance-management`, `employee-relations`, `learning-and-development`,
-`workforce-planning`, `employee-onboarding` (distinct from `revenue:activation`, which is users).
+**People depth.** Missing `performance-management`, `employee-relations`,
+`learning-and-development`, and `workforce-planning`.
-**Legal depth.** Missing: `ip-and-licensing`, `regulatory-compliance`, `audit-readiness`,
-`corporate-governance`.
+**Legal depth.** Missing `ip-and-licensing`, `regulatory-compliance`, and `audit-readiness`.
### Tier 2
-**Data & Analytics (CDO).** `demand-generation:marketing-analytics` covers marketing measurement
-only. No data governance, warehouse modeling, BI beyond marketing, or AI/ML governance — the last
-is increasingly a board-level obligation.
-
-**Customer Experience.** `revenue:retention` is the only adjacent skill. No support operations,
-escalation handling, or voice-of-customer program.
+**Communications and Investor Relations.** `marketing:public-relations` covers earned media. No
+internal communications, executive communications, crisis communications, or IR. IR only matters
+once there are investors.
-**Corporate Strategy / Corp Dev.** `executive:ai-research-analyst` does market research; nothing
-covers M&A, diligence, or scenario planning.
-
-**Operations depth.** Three skills. A business with physical production also needs
-`supply-chain-planning`, `quality-management`, and `capacity-planning`.
+**Product depth.** No dedicated `product-discovery`, `roadmap-prioritization`, or
+`pricing-experimentation` — the last overlapping `revenue:pricing-and-packaging`, so it may not
+warrant its own skill.
### Tier 3 — when the business needs them
- **Chief Sustainability / ESG** — emissions reporting and supply-chain diligence, increasingly
statutory rather than optional.
-- **Corporate Secretary** — currently folded into Administration.
+- **Corporate Secretary** — currently inside `legal-risk:corporate-governance`.
- **Internal Audit** — with a structural caveat: internal audit reports to the audit committee, not
the CEO. Placing it under `executive` would reproduce the independence failure it exists to
prevent. It needs a reviewer-class agent no chief can overrule.
-- **Investor Relations** — only once there are investors to relate to.
## Structural notes
-**Producer and auditor should not be the same agent.** The `executive:agent-hierarchy` skill makes
-this explicit, and the chart still partly violates it: most departments review their own work.
-`legal-risk` is the first genuine reviewer-class department. Security and Internal Audit are the
-next two, and both need the authority to block.
+**Reviewer independence is now modeled, not just described.** `security` and `legal-risk` each
+appear twice in `docs/AGENT-SURFACES.md`: once as a builder owning their own tree, once as a
+reviewer holding no write surface at all. The guard fails if a reviewer declares one, so the
+read-only property is structural rather than a promise.
-**Administration is a placeholder.** One charter, no specialists. It exists so orphaned
-responsibilities have somewhere to go rather than accumulating unowned.
+**Two departments report to the CMO.** `marketing` holds brand, content, and communications;
+`demand-generation` holds acquisition, conversion, and measurement. The split keeps specialization
+while halving what any one install loads.
-**How gap skills were scoped.** The `finance`, `people`, `legal-risk`, and `operations` skills were
-drafted against current senior job postings for those functions — FP&A leadership, commercial
-counsel, cybersecurity GRC, and HR business partner roles — so each skill's remit reflects what the
-role is actually accountable for rather than an assumption about it.
+**How gap skills are scoped.** Specialists in `finance`, `people`, `legal-risk`, `operations`, and
+`security` were drafted against current senior job postings for those functions, so each remit
+reflects what the role is actually accountable for rather than an assumption about it.
diff --git a/plugins/corporate-strategy/skills/chief-strategy-officer/SKILL.md b/plugins/corporate-strategy/skills/chief-strategy-officer/SKILL.md
index 587eb83..13f29fb 100644
--- a/plugins/corporate-strategy/skills/chief-strategy-officer/SKILL.md
+++ b/plugins/corporate-strategy/skills/chief-strategy-officer/SKILL.md
@@ -34,7 +34,9 @@ not strategy.
## What this role owns
-- The strategy of record and the choices in it.
+- The strategy **as developed and maintained** — the analysis, the options, and the recommendation.
+ Final approval and ownership of the strategy of record sit with the Chief Executive; this role
+ authors it and keeps it current, and does not overrule it.
- The portfolio view: which businesses get funded, held, or exited.
- Deal thesis and go/no-go on corporate development.
- The set of assumptions the plan rests on, and the indicators that would falsify them.
diff --git a/plugins/customer-experience/skills/voice-of-customer/SKILL.md b/plugins/customer-experience/skills/voice-of-customer/SKILL.md
index e87358b..3e91b50 100644
--- a/plugins/customer-experience/skills/voice-of-customer/SKILL.md
+++ b/plugins/customer-experience/skills/voice-of-customer/SKILL.md
@@ -10,8 +10,11 @@ loop is the whole value.
## Sources, weighted honestly
-- **Support contacts** — the highest-volume, least-biased source, and the most under-used. People
- contacting you have a real problem, unprompted.
+- **Support contacts** — the highest-volume and least *prompted* source, and the most under-used.
+ People contacting you have a real problem nobody asked them about. But the sample is strongly
+ self-selected: it excludes everyone who silently churned, worked around the problem, or would
+ never contact you. Treat it as operational evidence to be normalized per active account and
+ triangulated against churn and behavioural data — never as representative of the customer base.
- **Churn and loss reasons** — the most valuable and most under-sampled. People leaving have no
reason to be polite.
- **Interviews** — depth, small n, best for understanding *why* something in the data is happening.
diff --git a/plugins/data-analytics/skills/data-modeling/SKILL.md b/plugins/data-analytics/skills/data-modeling/SKILL.md
index 7ceb25c..d9c700e 100644
--- a/plugins/data-analytics/skills/data-modeling/SKILL.md
+++ b/plugins/data-analytics/skills/data-modeling/SKILL.md
@@ -9,8 +9,17 @@ description: Designs the warehouse and semantic layer — source-to-mart structu
Three layers, each with one job:
-1. **Raw** — source data, unmodified, append-only. Never transform on ingest; you cannot recover what
- you discarded, and you will want it.
+1. **Raw** — source data, append-only, otherwise unmodified. Do not apply *business* logic on
+ ingest: you cannot recover what you discarded, and the logic will need to change retroactively.
+
+ **Privacy and security transformations are the exception, and belong at ingest.** Credentials and
+ secrets should never land in the warehouse at all. Personal data that is not needed should be
+ dropped rather than stored and governed later, and identifiers you must keep but rarely need in
+ the clear should be tokenized or encrypted on arrival. Retention and deletion apply from ingest,
+ not from the marts.
+
+ The distinction: strip what you must not hold, keep everything you are entitled to hold, and
+ leave interpretation for later.
2. **Staging** — cleaned and conformed: consistent types, standardized names, deduplicated, no
business logic yet.
3. **Marts** — business-facing models shaped for how questions are asked.
diff --git a/scripts/build-readme.py b/scripts/build-readme.py
index a5890c9..2a5b80c 100755
--- a/scripts/build-readme.py
+++ b/scripts/build-readme.py
@@ -130,7 +130,32 @@ out += [
]
content = "\n".join(out)
+# The org chart's department table drifts the same way the README did. Generate it between
+# markers so the two cannot disagree; the analysis prose around it stays hand-written.
+chart_rows = "\n".join(
+ f"| `{d}` | {t} | {e} | {len(skills(d))}"
+ + (" · reviewer-class |" if d in REVIEWER else " |")
+ for d, t, e in ORDER
+)
+chart_block = (
+ "<!-- BEGIN GENERATED: departments -->\n"
+ f"| Department | Function | Executive | Skills |\n|---|---|---|---|\n{chart_rows}\n"
+ f"\n{len(ORDER)} departments, {total} skills.\n"
+ "<!-- END GENERATED: departments -->"
+)
+chart_path = "docs/org-chart.md"
+chart_current = open(chart_path, encoding="utf-8").read()
+chart_new = re.sub(
+ r"<!-- BEGIN GENERATED: departments -->.*?<!-- END GENERATED: departments -->",
+ lambda _: chart_block,
+ chart_current,
+ flags=re.S,
+)
+
if "--check" in sys.argv:
+ if chart_current != chart_new:
+ print(" docs/org-chart.md department table is stale — run: python3 scripts/build-readme.py")
+ sys.exit(1)
current = open("README.md", encoding="utf-8").read() if os.path.exists("README.md") else ""
if current != content:
print(" README.md is stale — run: python3 scripts/build-readme.py")
@@ -139,4 +164,5 @@ if "--check" in sys.argv:
sys.exit(0)
open("README.md", "w", encoding="utf-8").write(content)
-print(f"README.md regenerated — {len(ORDER)} departments, {total} skills")
+open(chart_path, "w", encoding="utf-8").write(chart_new)
+print(f"README.md and org-chart.md regenerated — {len(ORDER)} departments, {total} skills")