What problem does it solve? Keeping an agent's skill library current requires manually hunting through catalogs for capabilities that fill real gaps, then installing them safely. This Skill automates that discovery loop: it derives a concrete capability gap from repo signals, searches multiple skill catalogs, scores candidates against hard gates, and installs at most one trusted skill per run. ## Core Features & Use Cases - Gap-driven query derivation: Infers what to search for from failing skills in cron-state, open issues, non-template priorities, or recent log signals — and stays silent when no real gap exists. - Multi-catalog search with hard gates: Queries the npx skills CLI, curated GitHub indexes, and skills.sh, then filters candidates by gap fit, runtime compatibility, repo recency, and trust classification. - Trust-gated installation: Auto-installs only trusted, high-scoring candidates through bin/add-skill (updating skills.lock and aeon.yml), and routes untrusted matches to a manual-review notification instead. - Use Case: A scheduled run notices the rss-digest skill has failed three times, searches the catalogs for an RSS parser skill, finds a trusted well-maintained candidate, installs it on a branch, opens a PR, and notifies the operator. ## Quick Start Run the search-skill with an optional capability term like "rss" to find and install a gap-filling skill, or leave it empty to derive the gap automatically.
