CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

depth-estimation

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

Quality

68%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./skills/transformation/depth-estimation/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

68%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 lean, well-structured, and gives concrete protocol and setup guidance with real bundle-file references. Its main gap is the absence of validation/error-handling checkpoints in the continuous frame-processing workflow.

Suggestions

Add a validation/error-recovery step to the protocol workflow, e.g. an error event on model-load failure or frame-processing exception and a 'verify ready before transforming' checkpoint.

Make the script references explicit navigable links (e.g. 'See [scripts/transform.py](scripts/transform.py) for the full implementation') rather than inline imports only.

Provide a minimal concrete load_model/transform_frame body or point to a reference example so the interface template is copy-paste executable.

DimensionReasoningScore

Conciseness

The body is efficient — a compact hardware table, short bullets, and focused code/protocol blocks with no padding or explanation of concepts Claude already knows; only minor lines like 'near objects appear warm, far objects appear cool' could be trimmed.

4 / 5

Actionability

Provides concrete jsonl protocol examples, a working setup sequence, and a subclass template; the TransformSkillBase example uses '...' placeholder bodies, which is justified flexibility but leaves minor gaps, so it stops short of fully copy-paste-ready.

4 / 5

Workflow Clarity

The event sequence (setup → ready → transform → perf_stats → stop) is present, but this continuous/batch frame-processing skill shows no validation or error-recovery checkpoints (no error event, no model-load verification beyond 'ready'), so the batch-operation cap holds at 3.

3 / 5

Progressive Disclosure

Content is organized into clear sections (Hardware Backends, What You Get, Interface, Protocol, Setup) and references real bundle files (scripts/transform.py, transform_base.py), but those script references are inline rather than clearly signaled navigable links, keeping it just below 5.

4 / 5

Total

15

/

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 states a clear, specific purpose and niche but omits any explicit trigger guidance, which caps completeness at 3. Trigger terms lean technical and lack natural user phrasings.

Suggestions

Add an explicit 'Use when...' clause, e.g. 'Use when anonymizing camera feeds with depth maps or when the user asks for privacy-preserving depth estimation.'

Surface natural trigger terms users would say (e.g. 'depth estimation', 'anonymize video', 'blur identity') alongside the model/backend jargon.

Name a second concrete action beyond 'transforms' (e.g. 'generates colorized depth maps and overlays them on live feeds') to lift specificity.

DimensionReasoningScore

Specificity

Names the domain clearly ("Real-time depth map privacy transforms") and the toolchain ("Depth Anything v2 (CoreML + PyTorch)"), but the action word "transforms" is generic and only one concrete action is given, matching the 1-2 concrete-actions anchor.

3 / 5

Completeness

The "what" is clear (real-time depth map privacy transforms), but there is no "Use when..." clause or equivalent trigger guidance; per the rubric a missing when-clause caps completeness at 3.

3 / 5

Trigger Term Quality

"depth map" and "privacy" are relevant semi-natural keywords, but the heavy model/backend jargon ("Depth Anything v2", "CoreML", "PyTorch") is technical rather than what users say, and common variations like "depth estimation" or "anonymize" are missing.

3 / 5

Distinctiveness Conflict Risk

The specific model name and depth-map niche make it mostly distinct with minimal conflict risk, though "privacy transforms" could still overlap with other privacy/anonymization skills, keeping it just below 5.

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

skills/transformation/depth-estimation/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.