Blog

AI Development·August 31, 2026

GitHub's global Copilot model policy makes delegation a live choice

GitHub is enforcing its global Copilot model policy through September 1. Unconfigured models will inherit an enabled default unless administrators make explicit model-level choices.

AI Development·August 30, 2026

AWS Transform enters FedRAMP Class C scope in Ohio

AWS Transform is now inside the current FedRAMP Class C certification boundary in US East (Ohio). That changes service eligibility, not authorization of a complete modernization workload.

A row of empty white handled ceramic vessels waits in brass holders before a closed brass gate and an empty illuminated bay.

AI Development·August 30, 2026

GitHub Copilot self-serve seats become paid before access

GitHub is changing self-serve Copilot Business and Enterprise seat billing for credit-card and PayPal customers. Approve each assignment as a purchase and time removals before renewal.

Three parallel test-task bars contribute toward a dashed task-hour limit, with a note that work stops and discovered findings remain at the limit.

AI Development·August 30, 2026

AWS Security Agent can cap task-hours. The cap does not prove test coverage.

AWS Security Agent can now cap cumulative task-hours and revalidate selected findings. Use the cap as a spending ceiling, then choose focused or full retesting from the change scope.

AI Development·August 29, 2026

AWS puts Redshift skills inside AI agents—separate advice from authority

AWS now pairs Redshift-specific guidance with authenticated API execution in coding agents. Adopt it in stages so reference help, bounded queries, and migrations do not inherit the same authority.

AI Development·August 29, 2026

JFrog Traffic Controller moves package policy to the network edge

JFrog's new Traffic Controller reroutes public package downloads through Artifactory before developers or AI agents receive them. Test the network path before treating it as universal enforcement.

AI Development·August 29, 2026

AgentCore Memory can now enforce who gets which memories

AWS added JWT-authenticated Cedar authorization to AgentCore Memory. Teams can move tenant and user isolation out of scattered application checks, but the policy still needs adversarial verification.

AI Development·August 28, 2026

GitHub Copilot chat retention is moving from 28 days to the life of the account

GitHub will unify Copilot on the web, mobile, and cloud agent under one default-enabled policy. The same change extends github.com chat retention from 28 days to the life of the account.

AI Development·August 28, 2026

GitHub's Default code review setting moves to Balanced on September 28

GitHub will make Balanced the effective default for Copilot code review on September 28. Resolve inherited settings now, then verify effort and usage on real pull requests.

Industry Insights·August 28, 2026

Anthropic's Model Hardware Standard puts the safety boundary in the driver

Anthropic's MHS preview gives AI agents a common way to operate programmable lab and manufacturing equipment. Keep hard safety controls outside that software path.

A constructed request path shows an organization source policy narrowing into per-call domain and publication-date scope before eligible web results reach review.

Industry Insights·August 27, 2026

AgentCore Web Search can enforce source and date filters per call

AgentCore Web Search 1.2.0 adds server-side domain and publication-date filters to each request. Use admin policy as the ceiling, then test narrower task scope.

Four connected administration artifacts show aggregate evidence being read, one reversible target isolated for change, a person and policy shield at an approval checkpoint, and a separate scanner reading final workspace state.

Industry Insights·August 27, 2026

OpenAI’s Admin plugin turns chat into a change surface. Separate reads from writes.

OpenAI’s Admin plugin joins workspace analysis and supported admin actions in one conversation. Enable diagnosis first, then add narrow writes with verification.