bld-optimize-securitylisted

# bld-optimize-security — the final-boss security pass One skill that runs every *static* security tool we have against one app and returns a single triaged report. It **orchestrates** — it invokes other skills and checklists rather than re-deriving them. Detail lives in `references/`; this file is the conductor. Everything here is free and local. Dynamic runtime pentesting (actually attacking a live target) is deliberately **out of scope** — the tooling for it costs money, and BLD is free. ## What it combines | Source | Role | Cost | |--------|------|------| | Built-in `/security-review` skill | Fast first-pass vuln scan of the diff/code | free | | `gstack-cso` | Chief-Security-Officer deep review | free | | `gstack-review` | Pre-landing code review (security lens on the diff) | free | | `gstack-careful` | Guardrails around any destructive command this skill runs | free | | `gstack-investigate` | Root-cause a confirmed critical before fixing | free | | ECC `security-review.md` + `cloud-infrastructure-security.md` | Layered checklists (app + infra) | free (vendored markdown) | | ECC `production-audit.md` | "What breaks in prod" readiness (infra layers) | free (vendored) | | ECC `security-bounty-hunter.md` | Adversarial hunt for *remotely reachable* exploits | free (vendored) | | ECC `agent-architecture-audit.md` | **Only if the app has an LLM/agent** | free (vendored) | | `agentshield-checklist.md` | Audits our own `.claude/` harness config | free (markdown-only) | All ve