skills/npow/claude-skills/nap/Gen Agent Trust Hub

nap

Pass

Audited by Gen Agent Trust Hub on Oct 3, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes a local Python script located at ~/code/npow-dotfiles/scripts/nap.py. This script is part of the author's (npow) dotfiles repository and is used for local context management.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes project-related data files, such as MEMORY.md and journal entries, which could potentially contain malicious instructions if the data itself is untrusted.
  • Ingestion points: Reads MEMORY.md indexes andaimee journal files for compaction and archival purposes.
  • Boundary markers: None identified; the skill processes the files based on timestamps and file size without explicit data/instruction delimiters.
  • Capability inventory: The skill can delete files (pruning sagaflow directories), move files (archiving to subdirectories), and modify file content (compacting indexes) via the nap.py script.
  • Sanitization: There is no mention of content sanitization or validation before the script performs its compaction and archival logic.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 3, 2026, 08:42 PM