What is Hardening?
Hardening is the work of closing the gaps an attacker would use first. It covers unnecessary services, weak configurations, excess permissions, unpatched known vulnerabilities and fragile signing setups. RCMG does it for you or advises your team. It turns audit findings into changed systems, and it runs as one continuous step in the audit cycle.
Who this is for
This is for organizations that already know where they are exposed and need the exposure gone. Typical clients are mid-size and enterprise companies, exchanges and custodians, fund operators, and DAOs with significant treasuries.
Common triggers:
- An audit or penetration test produced a findings list, and your team can't close it fast enough.
- You are about to move treasury or client assets onto a new multisig or custody setup.
- Your board asked whether a Bybit-style signing compromise could happen to you.
- You are deploying AI agents with access to internal tools and data, and nobody owns the permission model.
- A regulator, insurer or institutional counterparty wants evidence that controls are in place, not only documented.
Why this matters now
The time between a public disclosure and an attack has collapsed. Median time to exploit is now minus 7 days, meaning exploitation typically starts before disclosure, compared with 771 days in 2018 (Mandiant M-Trends, 2026). Weaponization happens in under 24 hours (Microsoft DDR, 2026). In the first half of 2025, 32.1% of known exploited vulnerabilities were exploited on or before their CVE publication date (VulnCheck).
Attack automation has also become cheap. In research at UIUC, a GPT-4 agent exploited 87% of tested one-day vulnerabilities using only the CVE description. Around 72,000 CVEs are projected for 2026. In July 2026, JADEPUFFER became the first fully automated ransomware, and it got in through a known vulnerability on an exposed service.
Defenders are moving the other way. Median full patch time rose to 43 days from 32 (Verizon DBIR, 2026). Only 26% of critical KEV vulnerabilities are fully remediated. In complex enterprise applications, about 45% remain unpatched after a year (Qualys). The World Economic Forum (2026) found that 54% of organizations name the skills gap as their top security barrier.
What companies actually use it for
Closing exposed-service vulnerabilities
JADEPUFFER, like most opportunistic attacks, entered through a known flaw on an internet-facing service. Clients use hardening to shrink what is exposed, prioritize remediation against actively exploited vulnerabilities, and bring configurations back to an approved baseline. The goal is to close the specific doors automated tooling checks first.
Hardening signing and wallet infrastructure
The largest single crypto theft, Bybit's $1.46B in February 2026, came from a compromise of signing infrastructure rather than of the cryptography itself. Private-key compromises accounted for 43.8% of stolen crypto in 2024 (Chainalysis, 2026). Seventy-three percent of institutional crypto failures from 2022 to 2025 involved compromised wallet security (Chainalysis). Hardening here covers signer environments, quorum design, key storage and recovery. HSM integration cuts key compromise by 94%, yet only 34% of institutions test disaster recovery (BitGo, 2025).
Locking down identity and access
A single password without MFA led to a $3.09B breach (CyberVentures). Identity hardening addresses MFA coverage, privileged accounts, stale credentials and service accounts. These are the gaps that let a small foothold become a full compromise.
Constraining AI agents
Sixty-five percent of organizations had an AI-agent-caused incident in the past year (CSA/Token Security, 2026). Of those incidents, 61% involved sensitive data exposure. IBM found that 97% of organizations with AI-related breaches lacked proper access controls, and Salt Security found unauthorized or misconfigured MCP servers at 32.9% of organizations. Hardening scopes agent permissions to least privilege and closes tool and data paths the agents don't need.
What happens without it
Audit findings that sit unremediated become a list of known entry points. Given current exploitation timelines, a 43-day patch cycle leaves weeks in which attackers already have working exploits and your systems are still open. In the crypto context, $3.4B was stolen from platforms in 2025 (Chainalysis, 2026), and signing compromises are effectively irreversible once funds move. For AI deployments, Gartner expects 40% of enterprises to demote or decommission agents by 2027 over governance gaps. That outcome usually means sunk build cost and a stalled program.
What an engagement looks like
- Scope — We agree which systems, wallets, identities or agent deployments are in scope, and whether the engagement is done-for-you or advisory.
- Baseline and exposure review — We evaluate current configuration, access, patch state, signing architecture and recovery readiness against what is actually being exploited, not against a generic checklist.
- Prioritization — Findings are ranked by exploitability and business impact, so the riskiest gaps close first.
- Remediation — In done-for-you engagements, RCMG makes the changes with your team's approval. In advisory engagements, your team makes the changes with RCMG's direction and review.
- Verification and handoff — We confirm each change took effect and document the new baseline, which feeds into your next audit cycle.
What you receive — a hardened environment or a sequenced remediation plan, a documented security baseline, a record of what changed and why, residual-risk notes for anything deferred, and evidence suitable for boards, insurers or counterparties.
When this is not the right fit
- You haven't been assessed yet. If you don't know where you're exposed, start with an audit. Hardening works best on known findings.
- You need 24/7 monitoring and response. This is a scoped remediation engagement. It does not replace a managed detection or SOC service.
- Nobody can approve production changes. Without an internal owner who has authority to change systems, findings stay findings.
Frequently asked questions
How is this different from an audit? An audit tells you what is wrong, and hardening fixes it. The two run as a cycle: an audit finds the gaps, hardening closes them, and the next audit verifies the result and catches drift.
Should we choose done-for-you or advisory? Choose done-for-you if your team lacks the cap
