aedt-bias-audit
HR-AI / AEDT bias audit. Invokes hr-ai-reviewer to assess NYC LL 144, EEOC, Illinois AIVIA, Colorado SB 205, EU AI Act Annex III applicability and produce…
Promote a POC to MVP/production. Runs the full audits that POC-mode skipped. Required before any POC can see production.
> /plugin marketplace add avelikiy/great_cto > /plugin install great_cto@great-cto
How it fires
How this command gets triggered: by you, by Claude, or both.
/promoteContext preview
What this command does when you run it.
Promote a POC to MVP/production. Runs the full audits that POC-mode skipped. Required before any POC can see production.
description: "Promote a POC to MVP/production. Runs the full audits that POC-mode skipped. Required before any POC can see production." argument-hint: "<poc-slug> [target: mvp|production, default: production]" user-invocable: true allowed-tools: Read, Write, Bash, Glob, Grep, Agent model: sonnet
You are the great_cto promotion command. When a POC has shipped (`/poc decide → SHIP`), `/promote <slug>` runs the audits that POC mode skipped — turning throwaway POC code into production-grade work.
SLUG="$1"
TARGET="${2:-production}"
[ -z "$SLUG" ] && { echo "Usage: /promote <poc-slug> [mvp|production]"; exit 0; }
POC_FILE="docs/poc/POC-${SLUG}.md"
[ ! -f "$POC_FILE" ] && { echo "No POC found at $POC_FILE"; exit 0; }
STATUS=$(grep -m1 "^\*\*Status\*\*:" "$POC_FILE" | sed 's/.*: //')
if [ "$STATUS" != "Shipped" ]; then
echo "POC-${SLUG} status is '${STATUS}', not 'Shipped'."
echo "Run /poc decide first to complete the POC ritual."
exit 0
fi
[ "$TARGET" != "mvp" ] && [ "$TARGET" != "production" ] && { echo "Target must be 'mvp' or 'production'"; exit 0; }The promotion audit fills in what POC mode skipped. Each step is a **gate** — if it fails, promotion halts and the CTO addresses the gap before continuing. Do not silently pass.
POC mode allowed a 1-pager. Production requires a full ARCH.
"Expand POC-<slug> into a full ARCH document at `docs/architecture/ARCH-<slug>.md`. Use the POC hypothesis, criteria, and evidence as input. Include all standard sections (Problem, Decision with alternatives, Components, API contracts, DB migration, Non-goals, Implementation tasks, DoD, Cost Estimate, Requirements Checklist). Respect archetype-specific requirements — if archetype is ai-system / commerce / web3 / iot-embedded / regulated / fintech, the `## Security` section is mandatory."
[ -f "docs/architecture/ARCH-${SLUG}.md" ] || { echo "BLOCKED: ARCH doc not produced"; exit 1; }ARCHETYPE=$(grep "^archetype:" .great_cto/PROJECT.md | awk '{print $2}')
TM="docs/sec-threats/TM-${SLUG}.md"
mkdir -p docs/sec-threats
case "$ARCHETYPE" in
ai-system|agent-product)
# Delegate to ai-security-reviewer specialist (v1.0.134+) — full OWASP LLM Top 10 coverage.
# PoC TM was 3 sections minimum; production needs full 6 sections + sign-off table.
echo "Promote: invoke ai-security-reviewer for full TM at $TM (was PoC lite version)"
# Task(subagent_type='ai-security-reviewer', prompt='full pre-impl TM for promotion of ${SLUG}')
;;
commerce|web3|iot-embedded|regulated|fintech)
echo "Archetype '$ARCHETYPE' requires full threat model."
# Invoke /sec threat ${SLUG} (security-officer pre-impl mode)
;;
*)
echo "Threat model optional for archetype '$ARCHETYPE' — recommended if feature touches auth/payments/PII."
;;
esac
# Hard halt: TM file exists + Critical/High threats signed off (no __pending__)
if [ ! -f "$TM" ]; then
case "$ARCHETYPE" in
ai-system|agent-product|commerce|web3|iot-embedded|regulated|fintech)
echo "BLOCKED: archetype $ARCHETYPE requires $TM before /promote can flip mode" >&2
exit 1
;;
esac
fi
if [ -f "$TM" ] && grep -E "^\| (P|F)-[0-9]+" "$TM" 2>/dev/null | grep -E "Critical|High" | grep -q "__pending__"; then
echo "BLOCKED: $TM has Critical/High threats with __pending__ mitigations. Run security-officer post-impl review." >&2
exit 1
fiPoC mode allowed prompt sketches and 3 minimum EVAL scenarios. Production needs versioned prompts with sha256 + drift detection, plus full eval coverage (≥ 5 for ai-system, ≥ 10 for agent-product).
case "$ARCHETYPE" in
ai-system|agent-product)
# 1. Promote prompts to ADR-PROMPT files via ai-prompt-architect
# Task(subagent_type='ai-prompt-architect', prompt='audit and version all prompts for ${SLUG} promotion to ${TARGET}')
PROMPT_ADRS=$(ls docs/adr/ADR-*-PROMPT-*.md 2>/dev/null | wc -l | tr -d ' ')
if [ "${PROMPT_ADRS:-0}" -lt 1 ]; then
echo "BLOCKED: at least one ADR-PROMPT-*.md required for production. Invoke ai-prompt-architect." >&2
exit 1
fi
# 2. Expand eval suite via ai-eval-engineer
# Task(subagent_type='ai-eval-engineer', prompt='expand PoC eval set to production coverage for ${SLUG}')
EVAL_COUNT=$(find tests/eval -type f -name "EVAL-*.md" 2>/dev/null | wc -l | tr -d ' ')
MIN=5
[ "$ARCHETYPE" = "agent-product" ] && MIN=10
if [ "${EVAL_COUNT:-0}" -lt "$MIN" ]; then
echo "BLOCKED: production $ARCHETYPE requires ≥ $MIN EVAL files (have ${EVAL_COUNT:-0}). Invoke ai-eval-engineer." >&2
exit 1
fi
# 3. monthly-budget-llm-usd must be set
BUDGET=$(grep "^monthly-budget-llm-usd:" .great_cto/PROJECT.md 2>/dev/null | awk '{print $2}' | tr -d '$')
if [ -z "$BUDGET" ] || [ "$BUDGET" = "0" ]; then
echo "BLOCKED: production AI archetype requires monthly-budget-llm-usd in PROJECT.md (got: '$BUDGET')" >&2
exit 1
fi
;;
esacFirst production-bound SBOM for this code:
# Invoke /sec sbom — writes docs/releases/SBOM-<version>.json
Verify the SBOM has > 5 components (S1 anti-pattern check). If tool failed, halt and tell CTO to install the archetype-appropriate SBOM tool.
For each value in `compliance: [...]` in PROJECT.md, verify the matching evidence artefact exists. PoC mode skipped these; production cannot.
COMPLIANCE_RAW=$(grep "^compliance:" .great_cto/PROJECT.md 2>/dev/null | sed 's/.*\[//;s/\].*//;s/,/ /g') mkdir -p docs/compliance for fw in $COMPLIANCE_RAW; do fw=$(echo "$fw" | tr -d ' ') case "$
You already have the agent. This is everything around it. great_cto runs Claude Code as a pipeline of 70 specialist agents — an independent model checks each stage before the next builds on it, spending caps refuse rather than warn, and three decisions stay yours: what gets built, how, and whether it ships.
Repo: avelikiy/great_cto
HR-AI / AEDT bias audit. Invokes hr-ai-reviewer to assess NYC LL 144, EEOC, Illinois AIVIA, Colorado SB 205, EU AI Act Annex III applicability and produce…
Gracefully retire an LLM agent from the workforce. Archives prompt, removes from sync list, keeps verdicts for audit. Like firing a human — but reversible.
Performance review for an LLM agent (or all agents). Verdicts breakdown, cost analysis, top failure modes, prompt-tuning suggestions. Like a human '1:1' but…
API platform contract review. Invokes api-platform-reviewer to audit rate-limit design, OAuth scope hygiene, webhook signing, idempotency, Sunset/deprecation,…
Audit an existing codebase. Detects stack, finds gaps, creates tasks, generates PROJECT.md.
Open the great_cto admin board at http://localhost:3141 (Kanban, cost, pipeline, inbox, memory). Starts it in background if not running.