CtrlK
BlogDocsLog inGet started
Tessl Logo

converting-pytorch-to-litert

Converts PyTorch models (e.g. ResNet, timm, HuggingFace transformers) directly to LiteRT (.tflite) flatbuffer format. Use when converting PyTorch models to TFLite, setting up export environments, or troubleshooting torch-to-litert conversion bugs. Don't use for ONNX exports or converting existing TensorFlow models.

66

Quality

78%

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
SKILL.md
Quality
Evals
Security

Quality

Content

57%

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

A concise, well-structured set of conversion rules with concrete dependency guidance, but it stops short of showing the actual conversion step and omits validation, and the final code fence is unclosed.

Suggestions

Add the executable conversion step (e.g. the litert-torch convert call) so the core task is copy-paste ready, and close the open bash code fence.

Append a verification checkpoint to the workflow — e.g. loading the produced .tflite and checking supported ops — to catch conversion failures early.

Trim the "ATTENTION ALL AGENTS ... recent project findings" framing to reduce redundancy with the numbered rules.

DimensionReasoningScore

Conciseness

Mostly efficient and assumes Claude's competence (no basic-concept filler), but the "ATTENTION ALL AGENTS ... strictly follow these rules based on recent project findings" preamble and repeated emphasis on "direct/direct conversion" could be tightened.

2 / 3

Actionability

Provides an executable dependency command ("python3.11 -m pip install torch transformers timm litert-torch ai-edge-litert litert-cli-nightly") and a concrete error string, but the actual conversion code (e.g. the litert-torch conversion call) is missing and the bash fence is left unclosed at EOF.

2 / 3

Workflow Clarity

Numbered sections impose a loose ordered set of constraints, but there is no end-to-end conversion sequence and no validation/verification checkpoint (e.g. confirming the .tflite loads or ops are supported) for this multi-step conversion task.

2 / 3

Progressive Disclosure

At 26 lines for a single-purpose conversion skill with no bundle files present, the well-organized numbered sections meet the simple-skill bar for progressive disclosure without needing external references.

3 / 3

Total

9

/

12

Passed

Description

100%

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: concrete actions, natural trigger terms, explicit what-and-when guidance, and clear scope boundaries. Third-person voice is maintained throughout.

DimensionReasoningScore

Specificity

Lists multiple concrete actions — "Converts PyTorch models ... directly to LiteRT (.tflite) flatbuffer format", "setting up export environments", "troubleshooting ... conversion bugs" — and names concrete architectures (ResNet, timm, HuggingFace transformers).

3 / 3

Completeness

Explicitly answers both what ("Converts PyTorch models ... to LiteRT (.tflite) flatbuffer format") and when ("Use when converting PyTorch models to TFLite, setting up export environments, or troubleshooting torch-to-litert conversion bugs"), with added negative scope.

3 / 3

Trigger Term Quality

Covers natural terms users would say when needing this skill: PyTorch models, TFLite, LiteRT, torch-to-litert, ONNX exports, TensorFlow models, plus framework names (timm, HuggingFace transformers).

3 / 3

Distinctiveness Conflict Risk

Narrow niche (PyTorch → LiteRT direct conversion) with distinct triggers and explicit anti-triggers ("Don't use for ONNX exports or converting existing TensorFlow models"), making conflict with other skills unlikely.

3 / 3

Total

12

/

12

Passed

Validation

100%

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

Validation16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository
google-ai-edge/litert-torch
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.