BC operating-autodl-training
Operates remote model training jobs on AutoDL Linux servers over SSH. Use when starting a training run, checking whether training is still alive, reviewing GPU/CPU/memory/disk usage, reading recent logs, diagnosing abnormal interruptions, or summarizing the latest training outcome with next-step recommendations.
As a process C 56/100 · Has gaps — weak spots: result and completion, when it triggers, consistency
What is at stake
The skill contains fragments that, in the wrong hands, cost money or data. Below: what the installer risks and what the author should do.
The skill contains commands that delete files, rewrite disks or execute code fetched from the network. The agent may run them without asking if it believes the instructions require it.
Replace destructive commands with safe equivalents that ask for confirmation, scope them to one folder, and stop piping curl into a shell: pin a version and a checksum.
How to improve
- Address the high-severity findings: each costs 18 safety points. If one is a false positive, add the rule id to guard.allow in spec.yaml.
- 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 · 2
-
high Dangerous commands
cmd-encoded-execscripts/common.py:285Executes a base64/encoded payload (string literal in code, not executed)remote_command = f"printf %s {shell_quote(encoded_script)} | base64 -d | bash -s --"code literal
Medium and low: 1
-
low Dangerous commands
cmd-background-processscripts/remote_train.py:59Starts a background / autostarted process (string literal in code, not executed)nohup "$LAUNCHER_PATH" >> "$TRAIN_LOG" 2>&1 < /dev/null &
code literal
Files scanned: 13. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
✓ No remarks against the Agent Skills spec
Process rating: all ten parameters 56/100
- 0Result and completion. Does not say what the result is
- 20When it triggers. No condition that starts the skill
- 40Consistency. Frontmatter name (operating-autodl-training) differs from the folder (autodl-train)
- 50Failures and branches. 0 branches, has a failure section
- 60Tools and files. Uses tools (bash, python) that frontmatter does not declare
- 70Inputs and preconditions. Inputs and preconditions are listed
- 100Steps. 33 steps
- 100Execution cost. Instruction body is 1356 tokens
- 100Running it twice. No mutating operations
- 100Progress reporting. Reports progress
- medium Safety rules and hard prohibitions inside a skill: they belong in the system prompt, here they protect nothing
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 313: enough signal without eating the budget
- +4Structure: 13 headings
- +3Step-by-step instructions: 33 items
- +4Has examples (4 code blocks)
- +4Reference files are cited in the instructions (2 of 2)
- +3All 6 scripts are documented
Quality base 70; lint remarks subtract, signals add up to 100. Result: 91.