Opsera Expands Forge With Security and Governance for Legacy Apps
Opsera expands Forge with application discovery, security assessments and governance controls, adding an eight-dimension ForgeScore with NIST-aligned A–F grades as Anthropic reports 10,000+ critical flaws.
By Grace Kim
4 min read

What's News
- Opsera expanded Forge with application discovery, security intelligence, ForgeScore assessments and pipeline controls, available to Forge customers.
- ForgeScore uses eight dimensions with A–F grades and expanded support for the NIST Cybersecurity Framework; NIST does not certify or endorse such products.
- In a June update on Project Glasswing, Anthropic said partners using Claude Mythos Preview had found more than 10,000 high- or critical-severity flaws.
- Opsera says Forge works with Anthropic, OpenAI, GitHub Copilot and Cursor, and connects with more than 150 development, security and delivery tools.
- Forge runs alongside Opsera's Unified Insights platform, which tracks AI spending by model, tool and team.
By Grace Kim
Opsera is expanding its Forge platform with application discovery, security assessments and governance controls, widening an eight-dimension ForgeScore that tracks more closely to the NIST Cybersecurity Framework and hands leadership A–F grades on modernization risk. Opsera says the capabilities are available to Forge customers now.
The update lands against a swelling backlog of findable defects. In a June update on Project Glasswing, Anthropic said partners using Claude Mythos Preview had found more than 10,000 high- or critical-severity flaws, and named verification, disclosure and patching as a growing bottleneck.
AI writes code faster than most enterprises can map what they already run. An application can carry years of business rules, integrations and operational assumptions, and replacing one component without knowing those relationships turns a security fix into a production problem.
"Frontier AI has made finding vulnerabilities easier. Fixing them inside legacy code nobody understands is still expensive," said Vishnu Vasudevan, Chief Product Officer of Opsera. "With the Forge Software Factory acting as the AI harness around the models and agents teams already use, every modernization project becomes a security remediation project, and leadership gets a NIST-aligned grade they can take to the board."
What exactly changed inside Forge?
Four connected pieces of the modernization workflow now ship in the platform:
- Application discovery: Forge maps business rules and dependencies so teams can rank which systems to modernize first and see what a change will break before AI rewrites it.
- Security intelligence: Built-in checks flag known vulnerabilities, exposed credentials and outdated components, with explanations intended to tie technical findings to business risk.
- ForgeScore assessments: The eight-dimension scorecard expands its support for the NIST Cybersecurity Framework, with A–F grades designed to communicate modernization and security risk to leadership.
- Pipeline controls: Administrators define stages, approvals and goal scope to keep work inside agreed requirements and priorities.
Opsera also describes impact analysis for upgrades, which lets teams identify affected applications before a change ships. The link between discovery and security is straightforward: an outdated dependency matters more when engineers can see what relies on it and plan the replacement work.
The practical effect is that Forge moves upstream of code generation, into the stage where modernization projects usually fail.
What runs underneath the workflow?
Opsera's Forge product documentation describes a living knowledge graph connecting code, business rules, dependencies, data and policy. The point is to give people and coding agents a reusable understanding of an application rather than making every task rebuild its context from scratch.
That context feeds versioned specifications and work orders covering requirements, tests, architecture and completion criteria. The documented workflow applies policy and security checks before code generation, requires human approval for merges, and links changes back to their requirements and test evidence.
Opsera says Forge works with models and assistants from Anthropic, OpenAI, GitHub Copilot and Cursor, and connects with more than 150 development, security and delivery tools. The coding agent does the work inside a larger system that supplies context and controls how that work advances.
In plain terms, the pitch is context plus control, not another model.
What do the ForgeScore grades measure?
On its modernization website, Opsera describes ForgeScore as a baseline for comparing an application's condition before and after work. Its dimensions include the difficulty of changing tightly coupled systems, protection of trust boundaries, production resilience, clarity of business logic and readiness for future architectural changes.
That list explains why modernization readiness reaches past whether software compiles on a newer platform. A system can run fine while staying hard to understand, risky to change or dependent on fragile interfaces.
The expanded NIST alignment adds a shared risk vocabulary. The NIST Cybersecurity Framework organizes outcomes around Govern, Identify, Protect, Detect, Respond and Recover. The letter grades are Opsera's assessment mechanism; NIST does not certify or endorse CSF-related products or implementations.
The grades are management reporting, not an outside audit.
How does this connect to AI spending?
Forge sits alongside Opsera's Unified Insights platform, which connects AI usage and cost with engineering delivery. Opsera describes views of spending by model, tool and team, plus measures of how work moves from commits through review and to production.
The distinction matters for modernization budgets. More generated code or greater agent usage does not by itself show an application became safer or easier to maintain. A before-and-after assessment, a traceable change history and delivery data give leaders a firmer basis for judging what improved.
The open question is whether board-level grades and dependency maps, rather than faster code generation, will decide which legacy applications get modernized first.
As AI makes software transformation faster, that shared understanding becomes an increasingly important part of getting the work into production.
Source: unite.ai
More from Grace Kim
Show full bio
Market editor covering people, launches and day-to-day practice in entrepreneurship, startups and venture funding at Business Bearings.
2,206 articles