maintenance

Pass

Audited by Gen Agent Trust Hub on Feb 22, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION] (LOW): The skill utilizes the Bash tool to perform file maintenance tasks. It includes logic for counting lines and grepping date patterns in Plans.md and session-log.md. These operations are well-scoped to the skill's primary purpose of file organization.
  • [DATA_EXPOSURE] (LOW): The skill reads project metadata, logs, and planning files. Access is limited to internal project paths (e.g., .claude/memory/, Plans.md). No network-related commands or exfiltration patterns were identified.
  • [INDIRECT_PROMPT_INJECTION] (LOW): The skill processes untrusted data residing in Plans.md and session logs.
  • Ingestion points: Reads Plans.md and session-log.md via grep and wc.
  • Boundary markers: None present; the scripts parse the file content directly.
  • Capability inventory: The skill has Read, Write, Edit, and Bash permissions.
  • Sanitization: The logic uses standard Unix utilities (grep, awk) to extract specific fields like dates and status markers. While it processes user-generated content, the logic is descriptive and unlikely to lead to unintended command execution under normal circumstances.
  • [TIME_DELAYED_ATTACKS] (SAFE): While the skill uses date-based logic to trigger archiving (e.g., checking if a task is >7 days old), this is a core functional requirement and does not exhibit characteristics of a malicious logic bomb.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 22, 2026, 03:33 PM