Superpicky Cli

SuperPicky CLI skill: use absolute paths to scripts/install.sh and scripts/run.sh for automation; run.sh three entries (superpicky_cli, birdid_cli, --region-...

This skill is coherent with its stated purpose, but take these practical precautions before installing: 1) The included install.sh will git-clone an upstream repo (default: GitHub) and pip-install packages (including torch); review REPO_URL and the install.sh contents if you prefer a different source. 2) The installer creates a venv and writes files under $SKILL/.upstream and may download large model files if you pass --with-models — ensure you have disk space and bandwidth. 3) The runtime tools operate on your photo directories and implement a manifest-based reset that can delete temporary JPEGs generated by the program — review the manifest/reset behavior and test on a copy of your photos if you are cautious. 4) Pip installs pull code from PyPI (supply-chain risk inherent to Python packages); consider auditing dependencies or installing in an isolated environment. 5) If you want to avoid network clone, use --no-clone with a vetted local .upstream or set REPO_URL to a mirror you trust. Overall the package is internally consistent, but review the install/run scripts and the upstream code before granting it filesystem/network access.

Type: OpenClaw Skill Name: superpicky Version: 0.2.0 The SuperPicky CLI skill is a legitimate wrapper for an AI-powered bird photo culling and identification tool. It provides scripts for installation (`install.sh`), execution (`run.sh`), and a helper for fuzzy-searching eBird region codes (`ebird_region_query.py`). The skill manages an upstream repository (cloned from GitHub) and a local Python virtual environment. The instructions in `SKILL.md` are clear and focused on operational safety (e.g., using absolute paths). No evidence of data exfiltration, malicious persistence, or prompt injection was found; the tool's behavior is consistent with its stated purpose of processing and organizing bird photographs.