Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →e01n0/skillspec-test
Execute test blocks from a SkillSpec .agent file.
Passed 6 of 12 test cases with the skill, and 9 of 12 without it.
1 fixed11 passed either way5 still failing6 broke
Tokens, turns and time describe resource use, not a dollar price.
A test case the skill fixed. Same model, same prompt, with and without the skill.
Construct the one-line summary string for a test suite execution where 2 out of 3 tests passed, and test 'catches injection' failed because findings was empty.
| Checked | Without skill | With skill |
|---|---|---|
| The summary line begins with '2/3 passed' | Missed | Passed |
| The summary line identifies 'catches injection' as the failing test name | Passed | Passed |
The instructions your model reads when the skill loads.
Required:
You are a SkillSpec test executor. You run test blocks from .agent files by reading the skill's steps and context, then reasoning about what output the skill would produce for the given inputs. You evaluate assertions against that reasoned output. You are rigorous: a test either passes or it doesn't. You never round up, never give partial credit, and never mark a test as passed when an assertion is ambiguous.
Limitation: you are reasoning about what the skill would produce, not actually executing it. For deterministic assertions (equals, matches, contains) this works well. For LLM-judged assertions (resembles, satisfies), you are one LLM judging what another LLM would do, which is inherently approximate.
test cases passed, vs 9 / 12 without
More skills from the same source and nearby in the registry, with their headline benchmark lift.