k-skill-cleaner

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill uses npx to fetch updated instructions and run helpers via @nomadamas/k-skill@0. These commands point directly to the verified author's own package repository and represent intended vendor functionality.
  • [INDIRECT_PROMPT_INJECTION]: The skill parses local transcript files which may ingest untrusted data from historical conversations. 1. Ingestion points: log files at paths like ~/.claude/projects/**/*.jsonl. 2. Boundary markers: none present during regex log scanning. 3. Capability inventory: recommends deletion paths and removes files only upon user approval. 4. Sanitization: applies strict keyword and regex matching to count skill invocations, avoiding direct prompt execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 07:11 AM
Security Audit — agent-trust-hub — k-skill-cleaner