InstructFencePublic documentation
Safe, defensible prompt-injection testing guidance.
Start with InstructFence’s public workflow and security guidance for local AI agents and website chats. Detailed operational configuration remains in the authenticated private console.
Choose the right use case
Learn when to use the local AI agent, direct website chat, or retrieved-document chat workflow.
Prepare private prompts and compatible targets
Understand the private Prompt Library, target selection, and why capability compatibility matters before an evaluation.
Review preflight and coverage gaps
Understand immutable preflight, explicit authorization, request estimates, and why an unsupported check is a visible coverage gap.
Review findings and follow up
Learn the public concepts behind redacted findings, remediation guidance, strict report exports, and retesting.
Keep operations private
Target configurations, prompts, packs, runs, findings, evidence, and reports remain authenticated console data, not public documentation.
Frequently asked questions
Where should a new user start?
Start with the product overview, safe-use guidance, then create an account to access the private console.
Which concepts are covered?
The public guidance covers authorization, prompt-injection evaluation workflow, data boundaries, and result limitations.
How are public docs different from the console?
Public docs explain durable concepts; the console contains authenticated operational data and configuration.