v1.0.0

Memory Architect

BenignClawScan verdict for this skill. Analyzed May 1, 2026, 6:14 AM.

Analysis

This instruction-only skill is coherent for reorganizing existing memory files, but users should review it because it changes persistent agent memory and may structure sensitive personal or account-related notes.

GuidanceThis skill appears safe for its stated purpose, but it changes long-lived agent memory. Before installing or using it, be comfortable with the agent reading the full MEMORY.md, creating persistent memory files, and preserving existing directives. After use, review the generated files and consider keeping a backup.

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.

Abnormal behavior control

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.

Agent Goal Hijack

SeverityLowConfidenceHighStatusNote

SKILL.md

Preserves any system directives (NO_REPLY rules, heartbeat instructions)

The skill instructs the agent to carry forward existing memory directives that may affect future behavior. This is disclosed and likely intended to avoid losing existing rules, but those directives should be reviewed.

User impactOld or unwanted memory instructions could continue influencing future agent behavior after the restructuring.

RecommendationBefore or after running the skill, check preserved directives and remove any that should no longer guide the agent.

Cascading Failures

SeverityLowConfidenceHighStatusNote

SKILL.md

Replace MEMORY.md with a ~25-line index

The skill intentionally rewrites the main memory file and moves content into multiple tier files. This is purpose-aligned, but mistakes in classification or rewriting could persist across future sessions.

User impactImportant memory content could be misplaced or made less visible if the restructuring is wrong.

RecommendationKeep a backup or review a diff of MEMORY.md and the generated memory/ files after use.

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

SKILL.md

memory/ontology/graph.jsonl → Structured entities + relations (JSONL append-only)

The skill creates persistent structured memory that can be reused across tasks. This fits the memory-organization purpose, but persistent stored context can contain stale, sensitive, or misleading information.

User impactInformation from the old MEMORY.md may become easier for the agent to retrieve and reuse in later sessions.

RecommendationReview the generated memory files and graph, especially entries about people, projects, accounts, and instructions, and remove anything that should not persist.