From aria-ark
Use when working with VRS Health Check — validating Aria recording quality before MPS processing. Covers the run_vrs_health_check CLI, configurations, threshold checks, exit codes, overriding thresholds, and the Python API. Use whenever the user mentions VRS health check, recording validation, data quality checks, dropped frames, or pre-MPS validation.
How this skill is triggered — by the user, by Claude, or both
Slash command
/aria-ark:vrs-health-checkThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
VRS Health Check validates Aria recording quality — checking for dropped frames, sensor consistency, calibration validity, and timing issues. It runs automatically as part of MPS processing, but can also be run standalone to diagnose recording problems before submission.
VRS Health Check validates Aria recording quality — checking for dropped frames, sensor consistency, calibration validity, and timing issues. It runs automatically as part of MPS processing, but can also be run standalone to diagnose recording problems before submission.
This skill teaches concepts. For details:
run_vrs_health_check --helpfrom projectaria_vrs_health_check import vrs_health_check, then help(vrs_health_check)pip install projectaria-vrs-health-check (also installed automatically with projectaria-mps)run_vrs_health_checkMultiple preset configurations exist for different device generations and use cases (e.g. general-purpose vs SLAM-focused with stricter thresholds). Run --list-configurations to see all available presets, and --show-configuration-json <name> to inspect all checks and thresholds for a configuration.
Key concept: All available configurations are evaluated during every run regardless of which one you select with --choose-configuration. The selected configuration controls the exit code and console output, but the JSON output contains results for all configurations.
Two types of checks:
warn_threshold and fail_threshold values. Supports duration-dependent thresholds that relax for short recordings.Each check produces PASS, WARN, or FAIL. Use --show-configuration-json to discover all check names and their thresholds.
| Code | Meaning | MPS Implication |
|---|---|---|
0 | PASS | Good for MPS processing |
1 | WARN | Minor issues, MPS likely still works |
2 | FAIL | Fix before MPS processing |
64 | Usage error | CLI invocation problem |
66 | No input | File not found |
WARN guidance: Small numbers of dropped frames (3-5%) are marginal but usually processable. IMU warnings with camera pass usually means MPS will still succeed — MPS relies more heavily on cameras.
Two methods to customize thresholds — useful when default thresholds are too strict or too lenient for your use case:
--override-checks with the pattern "CheckGroup.metric.threshold_type=value"--override-check-file for complex, reusable custom configurationsBoth require --configuration-to-override to specify the base config. Use --show-configuration-json to discover valid check group and metric names. Checks can also be disabled entirely with ignore=true. See the customization docs for syntax and examples.
The health check is also available as a Python function for programmatic use and batch processing:
from projectaria_vrs_health_check import vrs_health_check
Run help(vrs_health_check.run_vrs_health_check) for the function signature and parameters.
VRS health check runs automatically as the first stage of MPS processing. A FAIL result blocks MPS from proceeding. Running health check standalone lets you diagnose and fix recording issues before uploading to MPS, saving time on the upload/processing cycle.
npx claudepluginhub facebookresearch/projectaria-plugins --plugin aria-arkProvides behavioral guidelines to reduce common LLM coding mistakes, focusing on simplicity, surgical changes, assumption surfacing, and verifiable success criteria.
Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.
Creates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.