Back to skill

Security audit

auto-report-generator

Security checks for vulnerabilities and agentic risk

Overview

The code and instructions largely implement the advertised report-generator, but there are several documentation/behavior mismatches and minor risks (undocumented API key handling, writable quota file in the home directory, and dangerous cleanup commands in docs) that should be clarified before trusting the skill.

This package appears to implement the advertised report generator, but before installing or running it consider: 1) The documentation, SKILL.md, and the actual CLI args differ — use scripts/generator.py's flags (--input, --api-key, --api-base, --no-ai, --sheet). 2) AI analysis requires you to provide an API key (via --api-key or environment in docs); be cautious which API base you point to — data summaries are sent to that endpoint. 3) The skill stores quota state at ~/.auto_report_generator/quota.json — expect persistent files in your home directory. 4) README/SKILLHUB include rm -rf commands pointing at /home/gem/... — do not run those blindly. 5) The docs mention paid token prefixes and tiers, but there is no remote billing/token-validation in the code; confirm how you will enforce or validate paid tiers if that matters. If you plan to use AI features, supply a trusted API key and API base; if you need guarantees that data never leaves your machine, run with --no-ai. If you want higher confidence, ask the author to: a) align SKILL.md/README with the actual CLI, b) remove unsafe example rm -rf commands, and c) document precisely how tokens/tiers are validated (or remove the misleading paid-tier text).

SkillSpector findings are pending for this release.