train-pose

Train/evaluate WiFi pose models honestly — camera-supervised (MediaPipe + CSI) and camera-free (WiFlow), always checked against the mean-pose baseline before any PCK is quoted.

Quality

74%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./harness/ruview/.claude/skills/train-pose/SKILL.md

Quality

Content

78%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 tight, honest instruction-only skill whose publish-time checklist is well sequenced and validated, but it stops short of being executable end-to-end: nothing tells Claude how to actually run the baseline, train the model, or invoke the claim check.

Suggestions

Add the concrete commands (or script paths) for running the mean-pose baseline, training each path, and invoking `ruview_claim_check`, so the workflow is executable rather than advisory.

Briefly define the three split types (chronological / blocked-gap / grouped-bucket) or point to where their construction is specified.

Add an explicit recovery step after the claim check — e.g. what to do when it flags an untagged or 100% claim — to complete the validate-fix-retry loop.

Dimension	Reasoning	Score
Conciseness	Lean and efficient — every line carries project-specific, non-obvious discipline (the retraction history, the ~50% mean-pose baseline fact, the leakage-free split taxonomy) and nothing re-explains concepts Claude already knows.	5 / 5
Actionability	The publish checklist is concrete (report delta in pp, run `ruview_claim_check`, tag MEASURED-EQUIVALENT only with the reproducer), but the core train/evaluate workflow has no commands or code, and key details like what `ruview_claim_check` is or how to construct each split are missing — matching the 'some concrete guidance but incomplete' anchor.	3 / 5
Workflow Clarity	The 'Before you publish a number' section is a clear 4-step sequence with an explicit validation checkpoint (step 3 flags untagged or perfect claims), matching anchor 4; it falls short of 5 because there is no error-recovery loop and the training/evaluation steps themselves are not sequenced.	4 / 5
Progressive Disclosure	At 29 lines with no need for external references (no references/, scripts/, or assets/ exist in the bundle), the three well-organized sections satisfy the rubric's under-50-lines exception for a top progressive-disclosure score.	5 / 5
	Total	17 / 20 Passed

Description

70%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 distinctive, concrete description that clearly states what the skill does in its project niche, but it entirely lacks a 'when to use this' trigger clause, which caps its completeness and weakens discovery by users who phrase the need differently.

Suggestions

Append an explicit trigger clause, e.g. 'Use when training or evaluating WiFi/CSI pose models, quoting PCK results, or comparing against mean-pose baselines.'

Add common synonyms users might say — 'pose estimation', 'wireless sensing', 'WiFi sensing' — to broaden natural trigger coverage.

Optionally mention the deliverable context (writeups, benchmark claims vs SOTA) so the 'when' covers evaluation-reporting situations too.

Dimension	Reasoning	Score
Specificity	Names several concrete actions — train/evaluate, camera-supervised (MediaPipe + CSI), camera-free (WiFlow), baseline check before quoting PCK — with only minor coverage gaps (no data-prep or inference detail), matching the 'several specific actions' anchor rather than the comprehensive 5.	4 / 5
Completeness	The 'what' is clear and concrete, but there is no 'Use when...' clause or equivalent explicit trigger guidance, which per the judging guidelines caps completeness at 3.	3 / 5
Trigger Term Quality	Good keyword coverage with terms a user in this project would naturally say ('WiFi pose models', 'MediaPipe', 'WiFlow', 'CSI', 'PCK'), but common synonyms like 'pose estimation' or 'wireless sensing' are missing, so it falls short of the comprehensive 5.	4 / 5
Distinctiveness Conflict Risk	A clear niche (WiFi/CSI pose modeling with named stacks like WiFlow and MediaPipe+CSI and a mean-pose baseline convention) makes it highly distinguishable with minimal conflict risk against other skills.	5 / 5
	Total	16 / 20 Passed

Validation

100%

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 — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: ruvnet/RuView
Path: harness/ruview/.claude/skills/train-pose/SKILL.md
Commit: 5ef001b

Reviewed: 3 days 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.