CD trtllm-model-onboard-multimodal
Onboard a HuggingFace multimodal model (vision/audio/video + text) to the TensorRT-LLM PyTorch backend. Use when writing a new `tensorrt_llm/_torch/models/modeling_{vlm}.py` plus its input processor and weight mapper, or extending an existing VLM.
This skill teaches you to onboard multimodal HuggingFace models into TensorRT-LLM on PyTorch. Promises a ready template for new `modeling_{vlm}.py` with input processor and weight mapper.
The description runs 12k tokens, but code quality scores 61, process score 35. Single file, no validation scripts. One medium-severity finding detected. No critical errors, though broken references flag documentation or code gaps. No sandbox testing or model runs performed.
Install if you already know TensorRT-LLM internals and can patch the template for your model. Works as a starting point, but don't expect it production-ready out of the box.
Onboard a HuggingFace multimodal model (vision/audio/video + text) to the TensorRT-LLM PyTorch backend.
As a process D 35/100 · Unfinished process — References files that are not bundled: references/mmmu.yaml, examples/llm-api/quickstart_multimodal.py
How to improve
- The SKILL.md body is over 5,000 tokens: move reference detail into references/ and load it when needed.
- The text references files that are not there: add them or drop the references.
- Your own cases (evals/evals.json, 4–6 real requests with expected answers): the full check would then run those instead of a model-drafted suite.
- A spec.yaml with trigger phrases and assertions — a behaviour contract for CI; `skilltest init` writes a template.
Guard findings · 1
✓ No critical or high findings
Medium and low: 1
-
low Secrets in code
secret-high-entropy-tokenSKILL.md:149High-entropy token-like string (may be an id, hash or a credential) (quoted — discussed, not commanded)**The lone existing exception** is `Qwen…del`, which keeps `Qwen…del` from `transformers` because Qwen2-VL is an outdated family on life support — not because passthro
quoted
Files scanned: 1. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
- warning
body-longSKILL.md body ≈ 12292 tokens (recommended < 5000); move details to references/ - warning
missing-refreference to a missing file: references/mmmu.yaml - warning
missing-refreference to a missing file: examples/llm-api/quickstart_multimodal.py - note
edit-residuethe text marks something as outdated (lines 50, 57, 138, 210, 329): check that old rules are not kept next to new ones — the full check reads the text for contradictions
Process rating: all ten parameters 35/100
- 0Tools and files. 2 referenced file(s) missing: references/mmmu.yaml, examples/llm-api/quickstart_multimodal.py
- 0Result and completion. Does not say what the result is
- 0Inputs and preconditions. Does not say what the process needs to start
- 20When it triggers. No condition that starts the skill
- 30Running it twice. 11 mutating operations with no state check
- 40Execution cost. Instruction body is 12292 tokens: crowds the task out of the window
- 60Steps. 86 steps, 4 vague phrases
- 100Failures and branches. 3 branches, has a failure section
- 100Consistency. Name and required fields are in place
- 100Progress reporting. Reports progress
- medium Safety rules and hard prohibitions inside a skill: they belong in the system prompt, here they protect nothing
- low The response is described with custom markup (7 tags): a typed call is more reliable
Everything here is measured from the skill text rather than judged by a model, so the numbers are checkable. A parameter weighs more when it is a more common reason for the process to stall.
Quality signals
- +5Description has no quoted example phrases that should trigger the skill
- +4Description does not say when NOT to use the skill (false activations)
- +3Output format is not stated: the model decides each time
- -5TODO / placeholder text left in the skill
- +2Single-language instructions
- +3Description length 247: enough signal without eating the budget
- +4Structure: 26 headings
- +3Step-by-step instructions: 86 items
- +4Has examples (7 code blocks)
- +1License stated
Quality base 70; lint remarks subtract, signals add up to 100. Result: 61.