asset-manager · Asset Manager

A skill for managing everything under outputs/: archive by date/platform/type, tag, search history, and generate asset inventories. It handles only what comes after creation — making past assets searchable and reusable.

Six operations: scan (refresh index), search (by keyword, platform, type, date, tags), archive (into date/platform/type folders), tag, list (asset inventory), and report (type/platform distribution plus output trends).

Example invocation: "Clean up and archive outputs/" or "Find the skincare cards I made before."

Full brief

Positioning

asset-manager is the housekeeper for outputs/ artifacts: post-creation asset management — archiving, tagging, search, statistics. Its job starts when content generation ends, keeping historical assets searchable and reusable instead of recreated.

Core capabilities

Six operations: scan (build/refresh index), search, archive, tag, list (inventory), report (statistics).
Multi-facet search: keyword, platform (Xiaohongshu/Weibo/Douyin/Zhihu/WeChat/X), type (image/video/text/card/poster/script), date, tags — combinable filters.
Structured archiving: files move to outputs/{date}/{platform}/{type}/; dates come from filenames first, mtime as fallback; platform/type inferred from filename and extension.
Safety guarantees: archive is dry-run by default, printing the plan; --apply moves only after confirmation; idempotent; CONFLICT on same-name-different-content targets — never overwrites.
Script-solidified: all side-effecting file operations run through scripts/assets.py (stdlib only). The LLM routes intent and presents results; it never hand-runs find/mv or writes JSON. Index and tags use atomic writes.
Statistics: type/platform distribution and 30-day output trends.

Workflow

Map user intent to a script subcommand (scan/search/archive/tag/list/report)
Refresh the index with scan if stale; search/list auto-scan when the index is missing
Run scripts/assets.py
Show the dry-run plan before archiving; apply only after confirmation
Present script output as Markdown tables

Inputs & outputs

Input	Notes
action	Required — scan / search / archive / tag / list / report
query	Optional — search keyword
path	Optional — target file or directory; defaults to `outputs/`
tags	Optional — comma-separated tags to add
filter_platform/type/date	Optional — platform / type / date filters
profile_name	Optional — filter by account profile

Output: Markdown tables — asset inventory (filename / type / source skill / date / tags), search results (file / match location / relevance / date), asset statistics (type / platform distribution / trends).

Boundaries with adjacent skills

Produce-layer skills (card, image-gen, video skills): generate content; this skill manages what comes after — archiving, search, tagging. It never produces assets, only manages them.

Fit

A cluttered outputs/ that needs structured archiving
Reusing past copy, cards, or scripts instead of starting over
Periodic asset reviews: per-platform distribution, 30-day output trends

Before you start

Specify the target directory (default outputs/); review the dry-run plan before applying an archive.
With a profile, search and archive scope narrows to the account and its platforms.
No API key or fees; the script is pure-stdlib Python.

asset-manager is part of the Aiglade Skill library. Invoke it from the Aiglade chat box in plain language.

← All 155 skills