The contained runner that drove the 2026-09 security audit, the README review and four rounds of design ideation lived only inside the ignored notebook. It moves to scripts/agent-runner/: the prepare, tools, audit, ideation and teardown scripts, the compose and container files, the allowlist proxy and the prompt templates, with the paths that assumed the notebook fixed and findings written to a caller-named directory. Run outputs, transcripts and the round-specific sheet scripts stay behind. docs/agent-runner.md states the method: the disposable git archive copy and the fail-closed proxy, how a task is shaped, union rather than intersection of findings across models, separate adjudication of every finding against the source, and the evidence a finding must carry.
16 lines
615 B
Plaintext
16 lines
615 B
Plaintext
# One POSIX-extended regex per line, matched against the requested hostname.
|
|
# tinyproxy denies any host that matches nothing here.
|
|
# If a run fails to reach a model, check `docker compose exec egress-proxy cat
|
|
# /tmp/tinyproxy.log` from this directory for the DENY line and add the host below.
|
|
(^|\.)deepseek\.com$
|
|
(^|\.)z\.ai$
|
|
(^|\.)bigmodel\.cn$
|
|
(^|\.)moonshot\.cn$
|
|
# Kimi For Coding plan endpoint is api.kimi.com/coding/v1 — NOT a moonshot host.
|
|
(^|\.)kimi\.com$
|
|
(^|\.)moonshot\.ai$
|
|
(^|\.)models\.dev$
|
|
(^|\.)opencode\.ai$
|
|
# Gemini (google provider) -> generativelanguage.googleapis.com
|
|
(^|\.)googleapis\.com$
|