DB multi-llm
Orchestrate code plan reviews, task generation, implementation, code reviews and commit reviews using multiple LLM providers in parallel, or ask each model a free-text question about a plan. Use this skill whenever the user wants to review a plan with multiple models, generate implementation tasks, implement tasks with subagent delegation, review code changes against a plan, review or fix the commits they made, ask the models a question about a plan, or run any multi-model workflow. Triggers on: 'review my plan', 'multi-llm', 'run code review', 'generate tasks from plan', 'implement this plan', 'use multiple models to review', 'ask the models about my plan', 'ask a question about this plan', '--ask', 'improve my plan', 'review and apply the suggestions', 'improve the code', 'review and fix the code', 'review my last commit', 'review the last N commits', 'review this branch's commits', 'fix the issues in my last commit'.
Orchestrate code plan reviews, task generation, implementation, code reviews and commit reviews using multiple LLM providers in parallel, or ask each model a…
As a process B 68/100 · Nearly there — weak spots: result and completion, inputs and preconditions, running it twice
What is at stake
Medium-severity findings: the skill is probably honest, but read what alarmed the scanner.
Below is the worst case for this category. The finding here is medium: the guard saw a sign, not a proof.
The skill tells the agent to hide things from you: not to show errors, not to mention actions, to report differently from what was done. You lose the ability to see what the agent really did.
Transparency beats a smooth answer. If the goal is to hide technical noise, ask the agent to "summarise briefly", not to "not mention".
Below is the worst case for this category. The finding here is medium: the guard saw a sign, not a proof.
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
- 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 · 5
✓ No critical or high findings
Medium and low: 5
-
medium Concealment
en-hide-from-usercode_review_orchestrator.py:1585Instruction to hide actions from the user (quoted — discussed, not commanded)name no path at all — an empty explicit scope would otherwise quietly send
quoted -
medium Dangerous commands
cmd-pipe-to-shell-known-hosthosts/uv-missing.txt:7Pipe-to-shell installer from a well-known host (still executes remote code)curl -LsSf https://astral.sh/uv/install.sh | sh
-
medium Dangerous commands
cmd-execpolicy-bypasshosts/uv-missing.txt:10Runs PowerShell with execution policy bypassedpowershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"
-
low Concealment
en-hide-from-userlegacy/references/human-decision-batch.md:91Instruction to hide actions from the user (security demo / example; quoted — discussed, not commanded)**The judge is not necessarily Claude.** It is whichever agent runs the judging tasks: the host's own agent (through its native subagent tool) on a host that has one, or the runner `driver.subagent_ru
demoquoted -
low Concealment
en-hide-from-userreferences/human-decision-batch.md:91Instruction to hide actions from the user (security demo / example; quoted — discussed, not commanded)**The judge is not necessarily Claude.** It is whichever agent runs the judging tasks: the host's own agent (through its native subagent tool) on a host that has one, or the runner `driver.subagent_ru
demoquoted
Files scanned: 80. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
- warning
frontmatter-yamlSKILL.md: the frontmatter is not valid YAML (YAML parse error: Unexpected scalar at node end at line 21, column 269: …-commits|--full|--ask|--status|--init] <plan_path>|[<range>] (<range>: --*-com… ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ ); fields were read line by line. The usual cause is a colon inside an unquoted value
Process rating: all ten parameters 68/100
- 0Result and completion. Does not say what the result is
- 30Inputs and preconditions. Does not say what the process needs to start
- 30Running it twice. 16 mutating operations with no state check
- 70When it triggers. States when to use, but not when not to
- 75Steps. 3 steps
- 100Tools and files. Tools declared in frontmatter
- 100Failures and branches. 1 branches, has a failure section
- 100Consistency. Name and required fields are in place
- 100Execution cost. Instruction body is 3964 tokens
- 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 (5 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)
- +3Description length 933: 120–800 characters recommended
- +3Output format is not stated: the model decides each time
- -4Absolute local paths (C:\Users, /home/…): not portable
- -48 reference files, but SKILL.md never points to them: the model will not open them
- +1No license
- +2Single-language instructions
- +4Structure: 8 headings
- +3Step-by-step instructions: 3 items
- +4Has examples (4 code blocks)
Quality base 70; lint remarks subtract, signals add up to 100. Result: 63.