Submit and review
Task intake, operator console, approval queue, run trace, changed-file view, and PR handoff page.
How AgentFoundry runs work
A product view over AI coding work that needs review.
An Evidence Report shows what an AI engineering run was asked to do, what it changed, which checks ran, what failed, what risks remain, what it cost, and what needs approval. AgentFoundry hides tool complexity behind one controlled way to run, inspect, and approve the work.
Control the work
Ownership, policy checks, identity, tool permissions, spend limits, and human approval decisions.
Run safely
Controlled execution paths, sandboxes, browser steps, and custom engineering agents behind one product interface.
Prove the result
Records, checks, screenshots, logs, diffs, risks, safe-stop notes, cost, and Evidence Report compilation.
Evidence lock
The page explains what teams can review and trust.
Product capabilitiesThis page explains what teams can use and trust, without exposing private implementation details.
Review controlsGovernance, execution, and evidence are explained through the work a reviewer can inspect.
Deployment privacyPrivate deployment topology, internal build paths, and confidential system mappings stay out of public architecture copy.
Execution workbench
Different coding tools stay behind one review process.
Execution hostsafe place to run work · shell records, policy checks, and run logs
Operator lanecoordination for longer tasks · memory, scheduling, browser actions, messages, and tool-call evidence
Implementation lanedeveloper-native coding path · terminal-first code edits, review notes, and focused diffs
Verification lanespecialist check path · patch generation, tests, review, and handoff evidence
Tool permission layerapproved tool access · typed tools, approvals, denial records, and audit trails
Work record API
Stores requirements, agent roles, dependencies, state changes, records, risks, and the next human decision.
Execution routing
Routes engineering work across execution hosts, verification paths, browser tasks, sandboxes, and custom agents.
Tool permissions
Treats repositories, issue trackers, CI, cloud systems, browser tasks, docs, typed tools, and internal APIs as approved capabilities with approval gates.
Evidence Report Compiler
Compiles evidence bundles from commands, diffs, test results, screenshots, logs, policy decisions, and safe-stop notes when runs are wired.
SDLC WorkGraph
The system of record for engineering-agent work: requirements, tasks, dependencies, tool calls, commands, changed files, risks, approvals, and review state.
Governance Core
The authority layer for agent identity, repo access, execution locks, policy gates, escalation, spend limits, and human approval.
Evidence Report
An Evidence Report shows what an AI engineering run was asked to do, what it changed, which checks ran, what failed, what risks remain, what it cost, and what needs approval.
Execution layer
AgentFoundry lets teams use different coding tools while keeping the same review, approval, and evidence process.
Tool permissions
Permissioned access to repos, issue trackers, CI, SAST/SCA, docs, browsers, APIs, cloud, governed tools, and approved internal systems.
Engineering Memory
Persistent repo context, policies, examples, report templates, owner decisions, and reusable engineering-agent lessons.