CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

depth-estimation

Real-time depth map privacy transforms using Depth Anything v2 (CoreML + PyTorch)

Quality

70%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Failed to scan

The risk profile of this skill

Fix and improve this skill with Tessl

tessl review fix ./SKILL.md

The canonical home for this skill is depth-estimation in SharpAI/DeepCamera

SKILL.md
Quality
Evals
Security

Quality

Content

75%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.

The body is efficient, well-structured, and actionable with concrete protocol and setup examples and an appropriately split implementation; the main gaps are a missing requirements.txt and the absence of explicit error/validation handling in the event loop.

Suggestions

Add the missing requirements.txt referenced by the setup block, or remove the reference if dependencies are documented elsewhere.

Inline-link the scripts (e.g. 'See [scripts/transform.py](scripts/transform.py)') so the implementation is discoverable from the body.

Document a brief error/recovery path in the protocol (e.g. what the skill emits on a failed frame or model-load error) to add a validation feedback loop.

DimensionReasoningScore

Conciseness

The body is lean: a hardware-backend table, a short feature bullet list, a subclass template, and concrete protocol/setup snippets, with no padded explanations of concepts Claude already knows; only minor lines (e.g. 'near warm, far cool') could be trimmed.

4 / 5

Actionability

Provides concrete, mostly executable guidance — JSONL protocol examples, a bash setup block, and a TransformSkillBase subclass template — with a minor gap: the referenced requirements.txt is not present in the bundle and the subclass example uses placeholder bodies.

4 / 5

Workflow Clarity

The stdin/stdout protocol clearly sequences the event loop (ready → transform → perf_stats, plus config-update and stop); no explicit validation checkpoints are documented, but as a streaming transform rather than a destructive/batch op the hard cap does not apply.

4 / 5

Progressive Disclosure

Well-organized into focused sections with the implementation appropriately split into scripts/transform.py (declared via the frontmatter capability and confirmed present); not a 5 because the body does not inline-link the scripts and references requirements.txt which is absent from the bundle.

4 / 5

Total

16

/

20

Passed

Description

53%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.

The description is specific and technically grounded with a clear niche, but lacks any 'when to use' trigger guidance and relies on somewhat technical phrasing rather than natural user keywords.

Suggestions

Add an explicit trigger clause, e.g. 'Use when anonymizing camera feeds, estimating scene depth, or privacy-blurring live video without revealing identity.'

Include natural synonyms users say ('depth estimation', 'distance map', '3D scene') alongside 'depth map'.

Name the concrete actions distinctly (e.g. 'generate colorized depth maps, overlay depth on live feeds, anonymize scenes') rather than the single verb 'transforms'.

DimensionReasoningScore

Specificity

Names the domain ('depth map privacy transforms') and one concrete action ('transforms') with specific tech (Depth Anything v2, CoreML, PyTorch), but does not list several distinct actions as the 4 anchor requires.

3 / 5

Completeness

It clearly states what the skill does ('Real-time depth map privacy transforms') but has no 'Use when...' clause or equivalent trigger guidance, which caps completeness at 3 per the rubric guideline.

3 / 5

Trigger Term Quality

'depth map', 'privacy', and 'real-time' are relevant keywords, but common synonyms users might say ('depth estimation', 'distance', '3D') are missing and the phrasing leans technical rather than natural.

3 / 5

Distinctiveness Conflict Risk

Naming the specific model (Depth Anything v2) and backends carves a distinct niche with only minor overlap risk against other privacy/transform skills; not a 5 because no explicit trigger phrases sharpen the boundary.

4 / 5

Total

13

/

20

Passed

Validation

93%

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

Validation — 15 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

frontmatter_unknown_keys

Unknown frontmatter key(s) found; consider removing or moving to metadata

Warning

Total

15

/

16

Passed

Repository

SharpAI/DeepCamera

Path

SKILL.md

Reviewed

Table of Contents

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.