Lexicon · observability

Runtime Receipt

Live proof that a tool or surface works right now, not that it worked at last deploy.

A runtime receipt is fresh evidence that a tool or surface actually responds correctly at this moment.

Why it matters

Code that passed CI is not the same as a surface that works in production. Agents act against live systems, and live systems drift, expire, and break under you. A runtime receipt is the difference between “should work” and “just confirmed it does,” which is the only version an agent should trust before acting.

In practice

Before trusting a surface, you pull a fresh receipt: call it, capture the response, confirm the shape. The receipt is dated and tied to the exact surface, so staleness becomes visible instead of assumed away.

Where it shows up in my work

In foundation3 I require a runtime receipt after edits to the live tool surface. The same discipline applies to any agent acting on systems it does not own.