Interactive Tool Inventory bootstrap. Use this when initializing a new project repo to configure the semantic tracking of Python/JS tools. It creates a dedicated RLM profile specifically for tools and performs the first intelligent distillation pass.
From tool-inventorynpx claudepluginhub richfrem/agent-plugins-skills --plugin tool-inventoryThis skill is limited to using the following tools:
acceptance-criteria.mdassets/resources/RLM_TOOL_STRATEGY_ANALYSIS.mdassets/resources/manifest-index.jsonassets/resources/prompts/rlm/rlm_summarize_tool.mdassets/resources/rlm_manifest.jsonevals/evals.jsonevals/results.tsvreferences/acceptance-criteria.mdrequirements.txtscripts/manage_tool_inventory.pyscripts/tool_inventory_init.pyDispatches parallel agents to independently tackle 2+ tasks like separate test failures or subsystems without shared state or dependencies.
Executes pre-written implementation plans: critically reviews, follows bite-sized steps exactly, runs verifications, tracks progress with checkpoints, uses git worktrees, stops on blockers.
Guides idea refinement into designs: explores context, asks questions one-by-one, proposes approaches, presents sections for approval, writes/review specs before coding.
This skill requires Python 3.8+ and standard library only. No external packages needed.
To install this skill's dependencies:
pip-compile ./requirements.in
pip install -r ./requirements.txt
See ./requirements.txt for the dependency lockfile (currently empty — standard library only).
Initialize the semantic Tool Inventory for a new project. This is the first-run workflow for tracking executable scripts.
Architecture Note: This skill delegates the actual data storage and generation to the
rlm-factoryplugin, but it strongly enforces a pre-configured schema specifically forplugins/andplugins/scripts.
Run the automated bootstrapping script. This script will ensure .agent/learning/rlm_profiles.json exists and will inject a tools profile if it doesn't.
python3 ./scripts/tool_inventory_init.py
The script above creates the target manifest, but YOU (the Agent) will execute the initial distillation pass if Ollama is unavailable, or you can delegate to batch mode if the project is massive.
Check what needs to be cached using the auditor:
# Hand off to the rlm-factory namespace
Trigger the 'rlm-curator' skill
If there are uncached tools:
Option A (Agent Distillation) - Recommended for < 20 tools: For each file identified as missing:
Option B (Batch Distillation) - Recommended for > 20 tools:
# Hand off to the rlm-factory namespace
Trigger the 'rlm-distill-agent' skill
Run the audit again to confirm 100% coverage:
# Hand off to the rlm-factory namespace
Trigger the 'rlm-curator' skill
manage_tool_inventory.py (inside the tool-inventory skill) to register single new scripts organically during development.