CtrlK
BlogDocsLog inGet started
Tessl Logo

correctness-reviewer

Always-on code-review persona. Reviews code for logic errors, edge cases, state management bugs, error propagation failures, and intent-vs-implementation mismatches.

65

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 ./.agents/skills/correctness-reviewer/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

100%

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

The content is a tightly-written, highly actionable reviewer persona: it specifies concrete defect patterns, gives precise confidence-calibration thresholds, defines what not to flag, and pins down an exact JSON output schema. It assumes Claude's competence throughout and wastes no tokens on basic explanations.

DimensionReasoningScore

Conciseness

The body is lean and assumes Claude's competence — it never explains what an off-by-one error or a race condition is, and every bullet delivers substantive, non-obvious guidance (concrete bug patterns, numeric confidence thresholds, what not to flag), matching the score-3 'every token earns its place' anchor rather than the padded level-1 or the slightly-loose level-2.

3 / 3

Actionability

For an instruction-only reviewer skill the guidance is concrete and specific: named defect classes with worked examples ('pagination that misses the final page when the total is an exact multiple of page size'), exact confidence thresholds (0.80+, 0.60-0.79, below 0.60), and a copy-paste-ready JSON output schema; per the scoring note, absence of code is not penalized when guidance is this actionable, so it clears level 3.

3 / 3

Workflow Clarity

This is a single-purpose reviewer skill whose process is clear and well-organized — trace execution, classify findings, calibrate confidence, suppress low-confidence findings, then emit JSON — with the 'suppress these' rule acting as an explicit quality checkpoint; the scoring note allows level 3 for simple skills when the content is clear and well-organized, and it is not level 2 because there is no missing sequence or validation gap.

3 / 3

Progressive Disclosure

The body is under 50 lines with no external references needed, and is organized into well-signaled sections ('What you're hunting for', 'Confidence calibration', 'What you don't flag', 'Output format'); per the scoring note this satisfies level 3 for short self-contained skills, and it is not level 2 because there is no inline content that should be split out.

3 / 3

Total

12

/

12

Passed

Description

60%

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 is specific and concrete about what the skill reviews, but lacks an explicit 'Use when...' trigger and leans on specialist jargon over natural user keywords, capping completeness and trigger-term quality. It is clearly written but reads more as an internal persona spec than a user-facing trigger description.

Suggestions

Add an explicit trigger clause, e.g. 'Use when reviewing code for correctness bugs, or when the user asks to find logic errors, edge cases, or off-by-one mistakes.'

Introduce more natural user-facing trigger terms ('find bugs in my code', 'check for edge cases', 'spot logic errors') alongside the specialist vocabulary.

Sharpen distinctiveness by contrasting scope with sibling reviewers, e.g. 'Correctness only — not style or performance.'

DimensionReasoningScore

Specificity

Lists multiple concrete actions: 'Reviews code for logic errors, edge cases, state management bugs, error propagation failures, and intent-vs-implementation mismatches' — matching the score-3 anchor of several specific concrete actions rather than the partial coverage at level 2.

3 / 3

Completeness

It clearly answers 'what' (reviews code for five named defect classes) but provides no explicit 'Use when...' trigger; per the judging guidelines a missing trigger clause caps completeness at 2, and it is not at level 1 because the 'what' is strong.

2 / 3

Trigger Term Quality

It has some natural terms a user might say ('code-review', 'logic errors', 'bugs', 'edge cases'), but the bulk is specialist vocabulary ('state management bugs', 'error propagation failures', 'intent-vs-implementation mismatches') and common variations are missing, matching 'some relevant keywords but missing common variations' rather than the full natural-term coverage at level 3.

2 / 3

Distinctiveness Conflict Risk

The correctness/logic-bug niche is somewhat specific and narrower than a generic 'code review', but the 'always-on code-review persona' framing and absence of distinct explicit triggers mean it could still overlap with sibling reviewer skills, matching 'somewhat specific but could still overlap' rather than the clear-niche level 3.

2 / 3

Total

9

/

12

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.

Validation13 / 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
udecode/plate
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.