Robotaxi Briefing
PassAudited by ClawScan on May 13, 2026.
Overview
This instruction-only news-briefing skill is coherent and low risk, but it will search public web sources and keep scoped local report files for its delta tracking.
This skill appears safe to install if you are comfortable with it making public web/news requests and saving/deleting its own timestamped report files under ~/.openclaw/workspace-group/memory. Review generated citations and make sure any referenced MEMORY.md rules are ones you trust.
Findings (2)
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.
The agent may contact Reddit, search providers, news sites, and resolved URLs while preparing the briefing.
The skill directs the agent to use external search and curl requests to gather and resolve public news URLs. This is expected for the stated briefing purpose, but it is still network/tool activity the user should be aware of.
Use `web_search` to search for... Also search X/Twitter: `site:x.com` ... `curl -sIL -o /dev/null -w "%{url_effective}\\n" "URL"`Use the skill only for public-source research, and review the cited links before relying on or redistributing the briefing.
Prior local report files and memory rules may affect what the agent treats as new or already covered, and matching old report files may be removed by the cleanup step.
The skill uses persistent local report files and references MEMORY.md rules to shape future output. This is aligned with delta tracking, but stored or externally maintained memory content can influence later briefings.
Reports are stored in `~/.openclaw/workspace-group/memory/robotaxi_report_*.md` ... Save the new report ... Then DELETE any `robotaxi_report_*.md` files older than 24 hours ... All rules in MEMORY.md Section 8 ... apply
Keep unrelated files out of the robotaxi_report_*.md pattern, and ensure any MEMORY.md rules used by the agent are trusted and current.