aiobotocore-bot

Use when addressing reviewer feedback on an aiobotocore PR — fetches every thread and top-level comment (including resolved), synthesizes into three buckets (asked / done / outstanding), then produces a per-thread action plan with one commit per group. Accepts `--focus=<databaseId>` to rank the triggering comment first and `--resolve` to mark threads resolved after a fix.

Use when applying the mechanical version + CHANGES.rst + pyproject.toml + uv.lock updates for a botocore sync (no-port = patch bump; port = minor bump, lower bound moves). Handles the Sphinx `^`-underline exact-length rule and drives `uv lock`. Caller commits the result.

Use when classifying a botocore version bump as no-port, port-required, or ambiguous. Diffs `$FROM..$TO` in a bare botocore clone, inspects every added/changed function in overridden files, and emits `CLASSIFICATION:` plus per-function reasons. Shared by the sync bot (to pick no-port vs port path) and the PR reviewer (to sanity-check sync-bot PRs).

Use when reviewing an aiobotocore PR that touches `aiobotocore/*.py` files with a botocore mirror. Compares each added line against the matching botocore function and emits `OVERRIDE_DRIFT:` (`clean` | `cosmetic-drift` | `behavioral-drift`) plus per-function detail. Distinguishes legitimate async gaps from unmatched additions that widen the sync diff.

Use at the end of every aiobotocore-bot workflow run to post the summary reply and swap the 👀 reaction for 👍. Handles the event-dependent target (inline-thread reply vs top-level PR/issue comment) and scopes reaction deletion to claude[bot] so human 👀 reactions aren't lost.