Expose only useful scanner context.
Agent tools can read coverage, attack surface, attack chains, findings, and fix prompts without downloading raw report noise.
Agent workflows
Give coding agents the scan context they need.
LaunchLock exposes a lightweight MCP-style endpoint so IDE and agent workflows can create scans, read findings, and pull focused fix prompts.
View manifestAgent contract
Manifest
Tools
Reports
Prompts
Report signals
ToolsScan and report reads
AuthAPI key headers
OutputStructured fix context
Tools
Scan and report reads
Auth
API key headers
Output
Structured fix context
Use the same API-key model.
MCP calls share the authenticated API key path already used by LaunchLock API routes.
Keep remediation grounded.
Prompts include scanner evidence and report context so the agent fixes what LaunchLock actually found.