You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
@@ -570,6 +571,11 @@ then asks whether to continue to the next report. Use
570
571
Use `--validate-reports-limit=<n>` to test the flow against a smaller number of
571
572
reports.
572
573
574
+
Some LLM providers or custom commands may incur token-based charges beyond a
575
+
regular subscription. The command asks for confirmation once before running those
576
+
providers. Batch runs with `--no-validate-reports-confirm` must also pass
577
+
`--llm-allow-paid-usage`.
578
+
573
579
#### LLM prompt and reasoning
574
580
575
581
The LLM prompt is designed to keep the model anchored to the Node.js threat
@@ -640,6 +646,7 @@ assessment.
640
646
|`--llm=none\|codex\|claude\|copilot`| Print prompts for manual LLM use or ask an LLM CLI to assess each report. Defaults to `none`. |
641
647
|`--llm-model=<model>`| Override the provider model and cache identity. |
642
648
|`--llm-command=<command>`| Override the command used for LLM assessment. The prompt is sent on stdin. |
649
+
|`--llm-allow-paid-usage`| Allow providers or custom commands that may incur token-based charges without prompting. Required for non-interactive paid-usage runs. |
643
650
|`--node-repo=<path>`| Path to a Node.js checkout containing `SECURITY.md` and `doc/`. Defaults to the current directory. |
0 commit comments