AB transformers-upgrade
Playbook for changing vLLM's Transformers version bounds. Use when bumping the pinned or maximum Transformers version, raising the minimum Transformers version, or deleting vendored configs, processors, or compat shims that upstream Transformers now provides.
Playbook for bumping Transformers versions in vLLM: handles pins, raises minimums, removes stale configs and shims. Six files, four scripts, 2177 tokens. Safety at 100, quality at 87—no critical or high findings, zero lint errors. No model runs recorded. Covers the standard dependency update workflow, but without sandbox testing it stays unclear how this plays out with actual model compatibility. The 63 process score signals gaps in documentation or automation.
Install if you're doing regular Transformers version maintenance on vLLM. For a one-off job, you'll finish faster by hand.
Playbook for changing vLLM's Transformers version bounds.
As a process B 68/100 · Nearly there — weak spots: result and completion, when it triggers, inputs and preconditions
How to improve
- 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 · 0
✓ No critical or high findings
Files scanned: 6. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
- note
edit-residuethe text marks something as outdated (lines 137, 155): check that old rules are not kept next to new ones — the full check reads the text for contradictions
Process rating: all ten parameters 68/100
- 0Inputs and preconditions. Does not say what the process needs to start
- 20When it triggers. No condition that starts the skill
- 30Running it twice. 17 mutating operations with no state check
- 40Result and completion. Does not say what the result is
- 100Tools and files. No external tools needed
- 100Steps. 26 steps
- 100Failures and branches. 1 branches, has a failure section
- 100Consistency. Name and required fields are in place
- 100Execution cost. Instruction body is 2550 tokens
- 100Progress reporting. Reports progress
- high The skill tells the model to perform an irreversible action with no human approval
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
- +1No license
- +2Single-language instructions
- +3Description length 259: enough signal without eating the budget
- +4Structure: 9 headings
- +3Step-by-step instructions: 26 items
- +4Has examples (3 code blocks)
- +3All 4 scripts are documented
Quality base 70; lint remarks subtract, signals add up to 100. Result: 87.
In the sandbox Скрипты не запустились
The skill's scripts were run in a throwaway machine: no network, fake keys in the home directory, a tracer watching. We wrote down what they did. Reaching for the network or for secrets caps the technical grade at C; a quiet run adds no points.
Запущено 4 скрипта; каждому дали двадцать секунд, поддельный домашний каталог с ключами и сеть, в которой ничего нет.
Ни один не дошёл до работы — им не хватило зависимостей, аргументов или файлов. Это не отзыв о поведении: наблюдать было не за чем.
scripts/diff_dumps.py | не запустился: ModuleNotFoundError: No module named 'regex' |
scripts/dump_configs.py | не запустился: ModuleNotFoundError: No module named 'vllm' |
scripts/dump_model.py | не запустился: ModuleNotFoundError: No module named 'torch' |
scripts/find_upstreamed.py | не запустился: ModuleNotFoundError: No module named 'regex' |
17 Oct 2026