CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

grilling

Grill the user relentlessly about a plan, decision, or idea. Use when the user wants to stress-test their thinking, or uses any 'grill' trigger phrases.

Quality

82%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Passed

No findings from the security scan

SKILL.md
Quality
Evals
Security

Quality

Content

93%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The body is a lean, highly actionable protocol: it defines a clear working model (design tree with a question frontier), an exact question format, a round-based interaction loop, and an explicit done condition with a user-confirmation gate. Its only gap is the absence of error-recovery guidance for messy sessions, which keeps workflow clarity at 4.

DimensionReasoningScore

Conciseness

The ~18-line body is dense with non-obvious instructions — the design-tree/frontier model, the round protocol, the facts-vs-decisions division of labor, and the termination condition — with zero padding and nothing that explains concepts Claude already knows. Every token earns its place, matching anchor 5; there is no trimmable over-explanation that would justify 4.

5 / 5

Actionability

The body gives a copy-paste-ready question format template (numbered question with title/body plus a recommended answer), an exact round protocol, and explicit rules for when to dispatch sub-agents and when to wait. For an instruction-only skill this is fully executable guidance covering the common case, matching anchor 5; it is not 4 because the format and protocol leave no gaps in how to act.

5 / 5

Workflow Clarity

The sequence is clear and carries explicit checkpoints: compute the frontier, ask numbered questions, wait for answers, recompute, and stop only when the frontier is empty — plus a confirmation gate before acting. It falls to anchor 4 rather than 5 because there is no error-recovery guidance (e.g., contradictory answers from the user, or a sub-agent returning conflicting facts), so the feedback loop is implicit rather than spelled out.

4 / 5

Progressive Disclosure

The skill is under 50 lines, has no bundle files, and inlines nothing that belongs in a separate reference — the whole body is the appropriate overview for its size. Per the rubric's simple-skills guidance, a short skill with no need for external references and well-ordered content scores 5; the prose flows logically from model to rounds to format to termination with no buried references.

5 / 5

Total

19

/

20

Passed

Description

62%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

The description answers both what and when in third person with a distinctive trigger verb, but it under-specifies both: the 'what' is a single action with no mechanics, and the 'when' defers to unenumerated 'grill trigger phrases' rather than listing natural trigger terms. It is serviceable but sits below the strong examples in the rubric.

Suggestions

Enumerate the natural trigger phrases the skill should fire on (e.g., "grill me on this", "poke holes in my plan", "play devil's advocate", "pressure-test this decision") instead of deferring to an undefined list of 'grill trigger phrases'.

Add one or two more concrete actions to the 'what' clause (e.g., "asks numbered questions in rounds and gives a recommended answer for each") to lift specificity from a single generic action.

Include common synonyms such as 'stress-test', 'pressure-test', and 'interrogate my plan' as explicit trigger terms to broaden natural keyword coverage.

DimensionReasoningScore

Specificity

"Grill the user relentlessly about a plan, decision, or idea" names the domain and one concrete action, but stops there — it never mentions the actual mechanics (numbered questions, recommended answers, rounds, design-tree mapping). Matches anchor 3 (domain plus 1-2 concrete actions, not comprehensive); it falls short of 4 because no list of specific actions is given, and exceeds 2 because the stated action is concrete rather than generic.

3 / 5

Completeness

Both parts are explicit: what ("Grill the user relentlessly about a plan, decision, or idea") and when ("Use when the user wants to stress-test their thinking, or uses any 'grill' trigger phrases"). Matches anchor 4 (both present, 'when' could be more specific) rather than 5, because the trigger guidance leans on an undefined set of 'grill trigger phrases' instead of concrete phrases.

4 / 5

Trigger Term Quality

Natural keywords "grill" and "stress-test" are present, but "uses any 'grill' trigger phrases" is self-referential — it defers to a phrase list it never enumerates, and common synonyms like "poke holes in my plan", "devil's advocate", "interrogate", or "pressure-test" are missing. Fits anchor 3 (some relevant keywords, missing common variations); not 4 because coverage is not good enough to call only 'a few' terms missing.

3 / 5

Distinctiveness Conflict Risk

The adversarial stress-testing niche and the distinctive verb "grill" give it a clear identity, but the object scope "a plan, decision, or idea" is broad and could overlap with planning, brainstorming, or decision-support skills. Fits anchor 4 (mostly distinct, minor overlap risk with closely related skills) rather than 5.

4 / 5

Total

14

/

20

Passed

Validation

100%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository

rpamis/comet

Path

eval/archives/grill-test-20261013/application/skills/grilling/SKILL.md

Reviewed

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.