Agent skill · data analytics · aiskillstore
home-bar-inventory-audit
Reconcile a recorded home-bar inventory against supplied physical bottle observations. Use after purchases, use, moves, or a long gap when every record and observation needs one review-only disposition.
Why this skill is useful
Adds a structured process for reconciling inventory discrepancies that the AI wouldn't reliably generate on its own.
What it needs
Requires python3 installed locally. About 2k tokens when loaded. Last updated 2026-08-07. 405 stars on the source repository.
What this skill does
Home-bar inventory audit Run a two-sided count. Every prior record and every physical observation must receive one disposition before the audit can finish. The skill produces a correction proposal. It does not edit an inventory system, infer what happened to a missing bottle, or decide freshness. Freeze the two snapshots Collect the current inventory record without changing it. Record each item with a stable record ID, exact label text, bottle size when known, fill-level scale, location, open date when known, and any prior uncertainty. Collect physical observations separately. Each observed bottle needs its own observation ID, label text, size, fill level, location, and direct notes. Unreadable labels, unknown fill, and missing dates stay unknown. A photograph does not become a verified field without user confirmation. Completion requires unique IDs for every record and every observation, plus an explicit value or unknown for each audit field. Propose one-to-one links Match an observation to a record only when identity evidence supports the same bottle. Similar category, flavor, or recipe role does not prove identity. Use one link state. matched means identity, location, and recorded facts reconcile. locationmismatch means identity matches while location differs. fillmismatch means identity matches while recorded and observed fill differ. needsreview preserves an uncertain identity or conflicting field. One record cannot link to two observations. An observation cannot satisfy two records. Completion requires a unique link for every proposed match and a note for every mismatch or uncertain link. Dispose every unlinked side An unlinked prior record receives missing or needsreview. Do not infer consumption, disposal, theft, or relocation. An unlinked physical observation receives added, duplicate, or needsreview. Use duplicate only when the user confirms that a separate identical bottle is intentional or redundant; category equivalence is insufficient. …
How to use it
Reference it in AdaL, Claude Code, Cursor or any coding agent — nothing to install:
@skills aiskillstore/home-bar-inventory-audit