OpenClaw Partner Guide
Analysis
This is a transparent operating guide for assistant behavior, with noteworthy but disclosed memory and scheduled follow-up practices.
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.
Checks for instructions or behavior that redirect the agent, misuse tools, execute unexpected code, cascade across systems, exploit user trust, or continue outside the intended task.
Create a real scheduled continuation immediately. ... If blocked again, schedule the next retry immediately. ... Never end with an unscheduled “later”.
The skill encourages durable scheduled follow-up across wait windows, which is a form of persistent agent behavior even though it is disclosed and bounded by task completion.
Checks for exposed credentials, poisoned memory or context, unclear communication boundaries, or sensitive data that could leave the user's control.
Use semantic retrieval before answering from prior decisions, dates, preferences, or todos. ... Store durable insights in long-term memory and raw events in daily logs.
The guide instructs the assistant to retrieve from and write to persistent memory/logs, which can affect future answers and may retain user context.