review-local command#
The local review command is available either as a script:
review-local README.md
or through the package entrypoint:
python -m moa review-local README.md
Synopsis:
usage: python -m moa review-local [-h] [--token TOKEN] [--save]
[--copilot-review] [--model MODEL]
[--prompt PROMPT] [-v]
file [file ...]
Reviews local files and prints markdown.
positional arguments:
file Local files to review
options:
-h, --help show this help message and exit
--token TOKEN GitHub personal access token. Resolution order: flag >
GITHUB_TOKEN env var > cached value
(/home/runner/.config/moa/review_pr.json) >
unauthenticated.
--save Save the resolved --token to
/home/runner/.config/moa/review_pr.json so it is used
automatically in future invocations. The file is created
with owner-only read permissions (0600).
--copilot-review Use GitHub Copilot (via GitHub Models API) to generate an
AI review. Requires a GitHub token with models access.
--model MODEL AI model used for --copilot-review. When omitted, Copilot
chooses the model automatically and falls back to
openai/gpt-4.1 if needed. Any model available on the
GitHub Models API is accepted.
--prompt PROMPT Add a follow-up prompt to the Copilot review session. The
prompt is sent as a continuation of the same conversation
so the model has full context from the initial review. Can
be used multiple times to ask several follow-up questions.
Only meaningful when --copilot-review is also set.
-v, --verbose Print progress information to stderr.
Examples:
review-local README.md
review-local --copilot-review --token "$GITHUB_TOKEN" moa/commands/review_pr.py
python -m moa review-local README.md
The command prints a # Local Files Review markdown report containing
the provided files and, if requested, a ## Copilot Review section.
It reuses the same token cache file as review-pr
(~/.config/moa/review_pr.json) and supports --save to persist
the resolved token. Use -v or --verbose to print progress
information to standard error while the command runs, including which
token source/type is used.
review-local supports the same --copilot-review, --model, and
--prompt flags as review-pr. Use --prompt one or more times
to ask follow-up questions as part of the same Copilot conversation
session:
review-local --copilot-review --token "$GITHUB_TOKEN" \
--prompt "Highlight any security concerns." \
moa/commands/review_pr.py