AC guardrails-safe-tools
Enforces guarded execution with safe_exec, safe_send, and safe_action. Use when a task may run shell commands, send channel messages, or call external APIs/actions that can mutate data or state.
As a process C 55/100 · Has gaps — weak spots: result and completion, consistency, running it twice
How to improve
For the model run — optional
- 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: 2. 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 55/100
- 0Result and completion. Does not say what the result is
- 0Progress reporting. Says nothing while it works
- 30Running it twice. 3 mutating operations with no state check
- 40Consistency. Frontmatter name (guardrails-safe-tools) differs from the folder (claw-guardirails-skill)
- 50Failures and branches. 0 branches, has a failure section
- 60Tools and files. Uses tools (bash) that frontmatter does not declare
- 70When it triggers. States when to use, but not when not to
- 70Inputs and preconditions. Inputs and preconditions are listed
- 85Steps. 18 steps, 1 vague phrases
- 100Execution cost. Instruction body is 493 tokens
- 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 194: enough signal without eating the budget
- +4Structure: 10 headings
- +3Step-by-step instructions: 18 items
- +4Has examples (3 code blocks)
Quality base 70; lint remarks subtract, signals add up to 100. Result: 84.
External checks
ClawHub: clean
This is an instruction-only safety skill that routes risky commands, messages, and external actions through guarded approval-aware tools.
LLM: benign (high) · VirusTotal: benign · 28 May 2026