v0.1.0

Agent Memory OS

BenignClawScan verdict for this skill. Analyzed May 1, 2026, 8:19 AM.

Analysis

This instruction-only skill is coherent and purpose-aligned, but it intentionally creates persistent agent memories that users should review to avoid retaining private or incorrect information.

GuidanceUse this skill if you want a structured long-term memory workflow. Before installing, decide where memories will be stored, keep secrets and private identifiers out of project/global memory, and require human review before lessons become canonical rules.

Findings (1)

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.

Sensitive data protection

Checks for exposed credentials, poisoned memory or context, unclear communication boundaries, or sensitive data that could leave the user's control.

Memory and Context Poisoning

SeverityLowConfidenceHighStatusNote

references/promotion.md

Canonical ... proven in multiple contexts ... strong enough to serve as a default rule or SOP.

The skill is designed to promote lessons into durable agent guidance, so inaccurate, private, or overly broad memories could affect later tasks if not reviewed.

User impactInformation saved into global or canonical memory could be reused later across projects, including mistakes or sensitive details if the user allows them in.

RecommendationReview and approve promotions into global or canonical memory, avoid storing secrets or personal data, and periodically prune stale or incorrect entries.