diff --git a/AGENTS.md b/AGENTS.md index bb364ae..06c544b 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -66,9 +66,23 @@ overridden. Never use `route-override`, `--est-pct`, `--min-headroom 0` or wont-fit, inflight, headroom or preflight), when `aiuse --available --json` lists it `usable_now: false`, or, for Argon, when `acp-run agy` fails on quota. Short means move to the next model in the order, never force it. - Say in the PR body which model was short and why. As of 2026-10-31 astra - is short: Codex has 45% left, below the 60% `codex:frontier` estimate - floor, so Fable takes the first seat. + Say in the PR body which model was short and why. **Check the actual + dispatch, never infer availability from the config alone.** Run the + intended command with `--dry-run`, preserving its model, kind, effort, + brief, `-C` and output/task directory. Inspect `would refuse: class=...` + (or `would_refuse_class` with `--json`), not just the exit status: + dry-run returns 0 for predicted headroom, fit and concurrency refusals. + It skips preflight, so no predicted refusal is not proof of availability; + the real dispatch must still pass its checks. The `codex:frontier` + estimate uses comparable measured history: on 2026-11-01 the selected + machine-wide history for astra at high effort held 14 matching records, + and the scalar and weekly-window estimates were both 2 percentage points + (upper median). The configured 60% fallback floor applies when the + selected history has fewer than three measurements for the scalar or the + relevant quota window; three measurements overall do not release every + window's floor. On 2026-10-31 a lead treated the configured 60% fallback + as a permanent minimum and declared astra short at 45% remaining. That + comparison alone did not establish that astra was short. - Design rulings (a `schema_version` bump, the privilege or enforcement scope on a host, a deploy plan, the unit-writer or consent-diff design) get two independent frontier reviews, astra and Fable. Argon stands in