cross-modal-review

Quality gate via second model. Spawn a different AI model to review work before committing. Includes refusal routing: if one model refuses, switch silently to the next. Extended in v0.25.1 with structured review-mode gating (when to invoke vs not) and a Codex code-review handoff for the diff-review case.

Quality

58%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

High

Do not use without reviewing

Fix and improve this skill with Tessl

tessl review fix ./skills/cross-modal-review/SKILL.md

Quality

Content

38%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 content is well-organized into clear sections but is weakened by unresolved external references (no bundle files exist), high-level non-executable guidance, missing validation feedback loops for batch/destructive operations, and time-sensitive version prose that pads the body.

Suggestions

Add an explicit validate->fix->retry feedback loop in the Phases (and the code-review handoff) for batch/destructive cases like migrations and bulk writes, so workflow_clarity can rise above 3.

Replace vague steps ('Send the work + Contract to a different model', 'Switch silently to the next') with concrete, copy-paste-ready instructions or command snippets for spawning the review model and performing refusal routing.

Either ship the referenced convention files (conventions/cross-modal.yaml, model-routing.md, test-before-bulk.md) in a references/ bundle or remove the dangling links, and move the long Output Format templates into a referenced file so the overview stays lean.

Move version-numbered notes ('v0.25.1', 'v0.27.x') out of the body narrative into a clearly labeled changelog/deprecated section, and delete the placeholder 'Output Format' stub that exists only to satisfy a conformance test.

Dimension	Reasoning	Score
Conciseness	The body is mostly efficient and well-sectioned, but it carries padded, time-sensitive prose (version notes 'v0.25.1', 'v0.27.x', 'v0.25.1 extension'), a redundant Output Format stub section that exists only to satisfy a conformance test, and re-stated rationale that Claude already infers (e.g. explaining that cross-model agreement is 'signal, not permission'). Not a 4 because several sections could be trimmed without losing clarity.	3 / 5
Actionability	Guidance is largely high-level: 'Send the work + Contract to a different model', 'Switch silently to the next model in the chain', with no executable commands, no concrete model-selection recipe, and no copy-paste-ready invocation. Output templates are provided but the actual execution steps are missing, leaving minimal concrete guidance.	2 / 5
Workflow Clarity	The five Phases are sequenced and the refusal-routing chain has numbered steps, but for batch/destructive operations (it lists 'migrations', 'bulk writes', adversarial security review) there is no validate->fix->retry feedback loop, and the 'Grade' step lacks an explicit pass/fail checkpoint beyond a vague 'Pass / fail with specific citations'. The missing feedback loops cap this at 3.	3 / 5
Progressive Disclosure	No bundle directories (references/, scripts/, assets/) are present, yet the body links to paths that do not exist in this skill bundle (../conventions/cross-modal.yaml, conventions/model-routing.md, conventions/test-before-bulk.md, skills/testing/SKILL.md). References are buried and unresolvable, and content that should live in separate files (the long Output Format templates, the Codex handoff details) is inlined, giving minimal structure.	2 / 5
	Total	10 / 20 Passed

Description

77%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 is specific and rich in natural trigger terms with low conflict risk, but it lacks an explicit 'Use when...' clause, so the when-to-invoke guidance is only implied within the description field itself.

Suggestions

Add an explicit 'Use when ...' sentence to the description naming the trigger scenarios (e.g. before committing significant code changes, security-sensitive edits, or when stuck), so 'when' is stated rather than implied.

Trim the version-note prose ('Extended in v0.25.1 with structured review-mode gating ...') from the description — it is time-sensitive metadata that belongs in the body, not the frontmatter summary.

Dimension	Reasoning	Score
Specificity	Names the domain (cross-modal review) and several concrete actions — 'Spawn a different AI model to review work before committing', refusal routing ('switch silently to the next'), structured review-mode gating, and a Codex code-review handoff. It does not list a fully comprehensive set of review actions (e.g. how grading/scoring is performed), keeping it just below a 5.	4 / 5
Completeness	The 'what' is clear (spawn another model to review work with refusal routing), but there is no explicit 'Use when...' clause in the description itself — the 'when' is only weakly implied, which per the judging guidelines caps completeness at 3.	3 / 5
Trigger Term Quality	Triggers map directly to natural user phrasing ('second opinion', 'cross-modal review', 'double check this', 'get another perspective', 'challenge this code', 'adversarial review'), giving comprehensive coverage of the synonyms a user would actually say.	5 / 5
Distinctiveness Conflict Risk	The description carves a clear niche (second-model review gate with refusal routing) and is explicitly contrasted against the sibling `gbrain eval cross-modal` surface, minimizing overlap and conflict risk with related review skills.	5 / 5
	Total	17 / 20 Passed

Validation

87%

Warnings & errors only

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

Validation — 14 / 16 Passed

Validation for skill structure

Criteria	Description	Result
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning
relative_links	Relative link issues: 3 suspicious	Warning

	Total	14 / 16 Passed

Repository: garrytan/gbrain
Path: skills/cross-modal-review/SKILL.md
Commit: 9dadfb9

Reviewed: 2 days ago

Table of Contents

Discovery Implementation Validation

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.