refresh-catalogs

Regenerate memory, diary, and skills catalog indexes via a dispatcher script.

Updated May 3, 2026
One-click install
npx skills add https://github.com/spikelab/multiplai-cc-mktplace --skill refresh-catalogs-spikelab
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: refresh-catalogs
Source: https://github.com/spikelab/multiplai-cc-mktplace/tree/main/plugins/multiplai-context/skills/refresh-catalogs
Command: npx skills add https://github.com/spikelab/multiplai-cc-mktplace --skill refresh-catalogs-spikelab

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires multiplai-core, claude-agent-sdk.

What problem does it solve? Catalog indexes used by the context manager for fast routing can become stale as memory, diary, and skills content changes. This Skill manually rebuilds those catalog indexes on demand, with state-aware skipping so unchanged sources are not regenerated unnecessarily. ## Core Features & Use Cases - State-aware regeneration: Rebuilds only catalogs whose source content changed since the last run, with memory and diary catalogs always processed. - Flexible flags: Supports --force to bypass skip logic, --dry-run to preview without side effects, and --only to selectively regenerate specific catalogs (memory, banks, diary, skills). - Isolated failure handling: A failing generator does not block others, and missing source directories or corrupt state files are handled gracefully. - Use Case: After editing several memory files, run the skill with --only memory to rebuild just the memory catalog so the context manager routes against fresh entries. ## Quick Start Ask the assistant to refresh the catalogs, optionally specifying flags such as a dry run or limiting regeneration to the memory and diary catalogs.

Frequently Asked Questions about refresh-catalogs

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I regenerate memory catalog indexes manually?

Run the refresh-catalogs skill, which invokes the generate_catalog.py dispatcher via uv run. By default it rebuilds all enabled catalogs using state-aware skipping, so only sources changed since the last run are regenerated.

How do I preview catalog regeneration without writing files?

Use the --dry-run flag. It reports which catalogs would be regenerated or skipped without writing files, modifying .generation-state.json, or making LLM calls, so there are no side effects.

Can I regenerate only specific catalogs like memory or diary?

Yes, pass --only with a comma-separated list of generator names such as memory, banks, diary, or skills. A generator named in --only runs even if its enable flag is off in the plugin configuration.

Why does the catalog script fail with ModuleNotFoundError for multiplai_core?

The script must be invoked with uv run --project pointing at the scripts directory, whose pyproject.toml declares the multiplai-core and claude-agent-sdk dependencies. Bare python or python3 lacks that environment and fails.

What does exit code 1 mean when regenerating catalogs?

Exit code 1 indicates partial errors, not total failure. Catalogs are still written for every source that succeeded, and re-running fills the gaps; a common cause is a 429 credits error on very large files.