Rollbar
PassAudited by ClawScan on May 1, 2026.
Overview
This Rollbar skill is transparent and purpose-aligned, but it needs a Rollbar token and can optionally change issue status, so users should use least-privilege credentials.
Install only if you want OpenClaw to access Rollbar. Start with a project-level read-only token, add write scope only when you intentionally want the agent to resolve, mute, or reopen items, and avoid storing the token in repository files.
Findings (3)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
A broader token could let the agent view multiple projects or change Rollbar item status if the user grants write scope.
The skill requires a Rollbar access token and supports both narrower project tokens and broader account/write-scoped tokens.
Use a token with `read` scope for monitoring; add `write` scope only if you need to resolve/mute items... account tokens grant broader access
Use a project-level read token for monitoring, add write scope only when you intentionally want status changes, and avoid account-level tokens unless multi-project access is necessary.
Accidental or overly broad use of write commands could mark active errors as resolved or muted, reducing alert visibility.
The helper script performs Rollbar API PATCH requests for item status changes such as resolve, mute, and activate.
api_patch "item/$ITEM_ID" '{"status":"resolved"}'Treat resolve, mute, and activate as approval-worthy actions, and prefer read-only tokens when using the skill only for monitoring.
If enabled, the agent will periodically query Rollbar and surface error data on the configured schedule.
The skill suggests user-configured periodic monitoring, which is disclosed and alert-focused rather than hidden persistence.
To get automatic alerts for new critical/error items, set up a cron job in OpenClaw... Recommended schedule: every 30–60 minutes during work hours.
Use cron only when desired, keep the schedule limited, and run scheduled monitoring with a read-only token where possible.