SECURITY REVIEW PROCESS
Use this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .csv, or .tsv file (e.g., adding columns, computing formulas, formatting, charting, cleaning messy data); create a new spreadshee
$ skillvault review xlsx
→ Fetching SKILL.md from https://github.com/anthropics/skills...
→ Running Stage 1 automated scan...
✓ Stage 1 PASSED — Risk score: 0/100
✓ Stage 2 PASSED — Human review complete
✓ Stage 3 PASSED — Sandbox execution clean
✓ Skill signed and published to registry
Stage 01
Automated Scan
Checked prompt injection, IoC matches, credential paths, shell commands, network calls. Score: 0/100
Stage 02
Human Review
Automated review. Risk: very_low. Score: 0/100.
Stage 03
Sandbox Execution
Skill executed in isolated container. Network traffic and file access logged. Zero deviations from declared behavior.
Stage 04
Cryptographic Signing
SHA-256 hash of skill directory signed via Sigstore keyless signing. Logged in public transparency log.
Full Scan Findings