Skip to main content
Version: v1.1.0
Executive visual. This page does not change TealTiger v1.1.0 contracts.

​
Promptware Kill Chain → Blast‑Radius Control (Executive View)

​
Key idea

Prompt injection is best understood as initial access in a multi‑stage “promptware” kill chain, not a single isolated bug. citeturn125search97 TealTiger’s job is blast‑radius control: constrain what the agent can do at runtime, even if earlier stages succeed. citeturn125search97
Think of TealTiger like a seatbelt and airbag for autonomous agents: it does not assume perfect prevention; it makes failures survivable.

​
One diagram (attack progression vs governance controls)

This diagram is intentionally high-level. It aligns with the article’s framing that prompt injection is only the first step in a broader operational chain. citeturn125search97

​
What execs should take away

  • Attackers don’t need perfect prompts; they need a path from input → action. citeturn125search97
  • “Prevent prompt injection” is not sufficient as a standalone strategy because the threat is multi-stage. citeturn125search97
  • Blast radius is the controllable variable: restrict tools, privileges, spend, and data movement.
  • TealTiger provides a deterministic decision boundary and audit-grade evidence to contain impact and prove outcomes.

​
Where this connects in TealTiger docs

  • Decision contract: /concepts/decision-model
  • Enforcement semantics: /architecture/enforcement-flow
  • Audit safety: /concepts/audit-and-redaction
  • Reason codes: /policy/reason-codes
  • Risk scores: /policy/risk-scores
  • Identity scoping: /concepts/execution-identity