rerun-lerobot

Ingest a LeRobot (HuggingFace) dataset into Rerun. Read when converting a LeRobot dataset to RRDs, splitting it into per-episode segments, or registering it on a Rerun catalog. Covers the built-in directory importer (log_file_from_path), the per-episode LeRobotReader, and when to drop to ParquetReader for custom control.

Quality

92%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Quality

Content

92%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A lean, well-structured ingestion guide with concrete code, explicit validation checkpoints, and a useful gotchas checklist, all backed by real bundle scripts. The only gap is a few undefined placeholder variables in the code snippets.

Suggestions

Define or inline the placeholder variables used in the code snippets (combined_rrd, rrd_dir, episode_parquet, and the zero_pad helper) so the examples are fully copy-paste-ready.

Add a one-line note showing how to obtain repo_id/dest (e.g. from the find script's output) so the Step 1 snapshot_download block runs end-to-end without inference.

Dimension	Reasoning	Score
Conciseness	The body is dense and information-rich, never explaining what LeRobot, Rerun, or parquet are, and every section (decision guidance, gotchas, code) earns its tokens; it assumes Claude's competence as the 5 anchor requires.	5 / 5
Actionability	Provides concrete, mostly copy-paste-ready code (snapshot_download + log_file_from_path, LeRobotReader streaming, RrdReader splitting, ParquetReader FK) and exact bash commands, but several snippets leave variables undefined (combined_rrd, rrd_dir, episode_parquet, zero_pad), the minor gaps that place it at 4 rather than 5.	4 / 5
Workflow Clarity	Steps 0–3 are clearly sequenced with explicit validation — the fetch script 'verifies that exactly the requested episodes landed' and 'asserts instead of leaving you to look', plus a 7-item Gotchas checklist and recovery guidance (start from an empty directory) — satisfying the explicit-validation and checklist requirements of the 5 anchor and avoiding the batch-operation cap of 3.	5 / 5
Progressive Disclosure	Well-structured into Step 0–3, Gotchas, and References, with one-level-deep, clearly signaled pointers to real bundle scripts (scripts/find_lerobot_dataset.py, scripts/fetch_lerobot_episodes.py both present) and sibling skills (rerun-chunk-processing, rerun-urdf, rerun-data-model), matching the clear-overview anchor.	5 / 5
	Total	19 / 20 Passed

Description

92%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong, specific description that names concrete actions and tools, gives explicit 'Read when' trigger guidance, and occupies a distinct niche. The only soft spot is trigger-term breadth — adding synonyms and the .rrd extension would round it out.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions — 'Ingest a LeRobot dataset into Rerun', 'converting ... to RRDs', 'splitting it into per-episode segments', 'registering it on a Rerun catalog' — and names specific tools (log_file_from_path, LeRobotReader, ParquetReader), giving comprehensive coverage rather than the minor gaps of a 4.	5 / 5
Completeness	Explicitly answers both what ('Ingest a LeRobot dataset into Rerun' plus the covered importers) and when ('Read when converting ... splitting ... or registering ...'), with concrete trigger phrases equivalent to a 'Use when' clause, matching the 5 anchor and avoiding the 3 cap.	5 / 5
Trigger Term Quality	Good natural-term coverage ('LeRobot dataset', 'Rerun', 'RRDs', 'per-episode segments', 'Rerun catalog', 'HuggingFace') that a user of this stack would actually say, but it lacks synonyms and explicit file extensions beyond 'RRDs', so it stops short of the comprehensive 5 anchor.	4 / 5
Distinctiveness Conflict Risk	Occupies a narrow, well-defined niche (LeRobot datasets into Rerun) with distinct tool names and triggers, so it is unlikely to fire for an unrelated skill — matching the 'clear niche with minimal conflict risk' anchor.	5 / 5
	Total	19 / 20 Passed

Validation

87%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 14 / 16 Passed

Validation for skill structure

Criteria	Description	Result
allowed_tools_field	'allowed-tools' contains unusual tool name(s)	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	14 / 16 Passed

Repository: rerun-io/rerun
Path: skills/rerun-lerobot/SKILL.md
Commit: 3efde05

Reviewed: 2 hours ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.