Long-Term Memory Manager
Analysis
The skill is coherent and local-only, but it can automatically write sensitive account or credential-related notes into persistent long-term memory, so it should be reviewed before use.
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.
MEMORY_FILE.write_text("\n".join(lines), encoding="utf-8") ... shutil.move(str(file_path), str(dest))The helper script can rewrite MEMORY.md and move daily memory files into an archive. This is purpose-aligned for memory maintenance, but it is still local data mutation.
Checks for exposed credentials, poisoned memory or context, unclear communication boundaries, or sensitive data that could leave the user's control.
r'(?:账号|account|credential):\s*(.+)' ... facts_section += entry ... write_memory_md(memory["sections"])
The script extracts any line marked as account or credential information from daily memory files and writes it into the persistent MEMORY.md file. The shown artifacts do not include redaction, approval, or validation before persistence.