evaluation

A practical, Bash-first guide to reliably testing probabilistic AI agents from the Linux terminal using curl and jq: define JSONL test cases, run a tiny harness against your endpoint or a deterministic stub, assert via contains/regex, manage golden files, run in parallel, and prep for CI—while constraining randomness (temp 0, seed), normalizing outputs, mocking tools/time, and keeping fast, reproducible smoke tests.

A practical guide to building reliability-first AI agents on Linux with Bash and standard tools. It explains common failure modes (probabilistic AI, flaky networks, crashes, low visibility) and offers copy-paste patterns: hardened entrypoint, timeouts/retries/backoff, idempotency/state, structured JSON logs, and systemd sandboxing/self-heal—plus a ready-to-run inbox-to-outbox summarizer using curl/jq and ops tips.