CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

arena

Run N parallel candidates at one brief and synthesize: rubric-first framing, isolated fan-out, a cross-judge on a different model, criterion-by-criterion pick, hand grafting. When one attempt would lock in the wrong shape.

Quality

80%

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

Fix and improve this skill with Tessl

tessl review fix ./skills/arena/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

88%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 an exemplary lean orchestration skill: a clear A–F phase sequence with explicit verification, error-recovery loops, dropout handling, and tie-break rules, all written for a competent reader with no padding. The only soft spot is the long multi-provider routing paragraph inlined in Phase A, which is the sole candidate for extraction to a reference file.

DimensionReasoningScore

Conciseness

The body is lean throughout and explains only operational specifics Claude would not know, e.g. "an unset mode resolves to `disabled` and all candidates would share one checkout". Even the intro rationale ("agreement across independent attempts is high-signal, and divergence exposes what the brief under-specified") directly drives Phase E behavior, so every token earns its place — anchor 5, above 4 (no trimmable over-explanation found).

5 / 5

Actionability

Provides a copy-paste command (`OMC_TEAM_WORKTREE_MODE=detached omc team --no-decompose 1:claude,1:codex,1:gemini "<brief>"`), concrete config paths (`.claude/omc.jsonc`, `omc config`), and explicit failure-handling branches (dropout replacement, cross-family fallback). Some directives remain abstract ("Fold grafts in by hand so the result stays coherent"), matching anchor 4 — mostly executable with minor gaps — below 5.

4 / 5

Workflow Clarity

Phases A–F are clearly sequenced with explicit validation and feedback loops: Phase F requires verification under "real command output, fresh evidence. Do not paper over it", and failures route backward ("either the rubric was wrong (reframe and rerun) or a graft was missed (back to Phase E)"). Phase D adds tie-break rules and Phase B dropout handling — anchor 5.

5 / 5

Progressive Disclosure

A well-sectioned, self-contained single file (~107 lines) with no nested or buried references; external skill references (`verify`, `execute`, `research`) are clearly signaled. The dense multi-provider routing block in Phase A step 3 is the one stretch of config detail inlined where a separate reference file could sit — anchor 4, above 3 but short of the ideal split.

4 / 5

Total

18

/

20

Passed

Description

67%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 clearly states what the skill does through a dense list of named phase-actions and includes an explicit when-clause, keeping it distinct from neighboring skills. Its main weakness is trigger-term coverage: the natural phrases a user would say (bakeoff, compare approaches, try a few ways) live in the body rather than the description, and the compressed jargon slightly blurs capability specificity.

Suggestions

Add natural trigger phrases to the description itself, e.g. "Use when the user asks for a bakeoff, a comparison of approaches, or to try a few ways and pick" — these phrases currently exist only in the body's When-to-use section.

Broaden the when-clause beyond the single scenario "When one attempt would lock in the wrong shape" to also cover judgment-sensitive work (migration strategy, API surface, data structure choice), which the body already names.

Unpack the telegraphic action list slightly — replace jargon like "isolated fan-out" and "hand grafting" with plain concrete actions (e.g. "dispatch isolated parallel attempts", "port the best pieces by hand") so each capability is unambiguous.

DimensionReasoningScore

Specificity

Lists several concrete actions — "Run N parallel candidates at one brief", "a cross-judge on a different model", "criterion-by-criterion pick", "hand grafting" — but the telegraphic listing leaves actions like "hand grafting" and "isolated fan-out" underspecified to a reader. Fits anchor 4 (several specific actions, minor gaps in coverage), not 5 (comprehensive, fully concrete) and clearly above 3 (more than 1-2 actions named).

4 / 5

Completeness

The "what" is explicit ("Run N parallel candidates at one brief and synthesize") and a "when" clause is present ("When one attempt would lock in the wrong shape"), so it is not capped at 3. However, the when-clause covers only one abstract scenario without concrete trigger phrases, matching anchor 4 ('when' could be more explicit or specific) rather than 5.

4 / 5

Trigger Term Quality

"When one attempt would lock in the wrong shape" is the only user-voiced trigger phrase; the natural terms users would actually say for this skill — "bakeoff", "comparison of approaches", "try a few ways and pick" — appear only in the body's When-to-use section, not the description. Matches anchor 3 (some relevant keywords but missing common variations), below 4 (several natural terms are absent, not just a few).

3 / 5

Distinctiveness Conflict Risk

"N parallel candidates at one brief" carves a clear niche distinct from single-run execute/research skills, and terms like "cross-judge" and "grafting" are distinctive. Minor overlap risk remains with general orchestration/team-dispatch skills, so it fits anchor 4 (mostly distinct, minor overlap) rather than 5 (clear niche with fully distinct triggers).

4 / 5

Total

15

/

20

Passed

Validation

81%

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

Validation — 13 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

metadata_version

'metadata.version' is missing

Warning

metadata_field

'metadata' should map string keys to string values

Warning

frontmatter_unknown_keys

Unknown frontmatter key(s) found; consider removing or moving to metadata

Warning

Total

13

/

16

Passed

Repository

Yeachan-Heo/oh-my-claudecode

Path

skills/arena/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.