CtrlK
BlogDocsLog inGet started
Tessl Logo

model-right-sizer-audit

One-shot, PER-CALL-SITE audit of a repo's EXISTING LLM calls — every place code invokes a model (an SDK/API call, a sub-agent dispatch, an agent frontmatter definition), decomposed by INTENT into the distinct jobs it does, never a grep hit on a model name and never one candidate per file. A skill pinned by one static `model:` key still gets its steps split by intent when severable under Claude Code's per-turn model binding. DELEGATES each call to `model-right-sizer-dryrun` (never re-implements its scoring) and merges results into ONE schema-conformant JSON blueprint committed at the TARGET REPO'S ROOT via a PR — never a markdown table. Distinct from `model-right-sizer-dryrun` (invoke directly for one hypothetical task) and `model-right-sizer-install` (the standing before/after mandate). Use when someone says "audit this repo's model calls", "right-size the models in <repo> per call site", "find every LLM call and right-size it", or "commit a model right-sizing blueprint for <repo>".

70

Quality

88%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

SecuritybySnyk

Low

Low-risk findings worth noting

SKILL.md
Quality
Evals
Security

Quality

Content

77%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 a thorough, well-sequenced workflow with strong actionability and explicit validation checkpoints, but it is markedly verbose and inlines a great deal of rationale that could live in separate reference files. Tightening the prose and pulling the long decomposition/severability and cleanup-contract discussions into references would lift both conciseness and progressive disclosure.

Suggestions

Move the extended decomposition/severability rationale and the full cleanup-contract discussion into reference files, leaving SKILL.md as a concise overview that links one level deep — this would improve both conciseness and progressive disclosure.

Cut the narrative 'an earlier version did X, which let Y through' passages down to the resulting rule; keep the guardrail, drop the post-mortem, to reduce verbosity.

Eliminate the repeated restatements of the cleanup contract across steps — point back to the single canonical section instead of re-explaining it each time.

DimensionReasoningScore

Conciseness

The ~840-line body is mostly efficient expert guidance, but it carries notable padding — long discursive paragraphs explaining why each guardrail exists, repeated restatements of the cleanup contract across steps, and narrative accounts of prior-version mistakes — that could be tightened without losing clarity.

3 / 5

Actionability

Concrete, executable shell snippets (git/gh plumbing, validation via uv run, heredoc-assembled PR body) cover the common cases and are largely copy-paste ready, with only minor placeholders like <marketplace-checkout> and {N}/{M}/{K}/{R} left for the operator.

4 / 5

Workflow Clarity

A clear 7-step sequence with explicit validation checkpoints (validate the merged blueprint before writing, poll gh pr checks to a terminal state, per-exit cleanup contract) and real feedback loops (fix-and-re-poll), satisfying the destructive/batch validation requirement rather than tripping the cap-at-3 rule.

5 / 5

Progressive Disclosure

Two real, correctly-referenced scripts (render_pr_table.py primary, render_pin_audit.py legacy) and the schema are signaled one level deep, but the bulk of the decomposition/cleanup/severability rationale is inlined monolithically into SKILL.md rather than split into reference files, so structure is present but not well partitioned.

3 / 5

Total

15

/

20

Passed

Description

100%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 comprehensive and concrete: it states exactly what the skill does (per-call-site intent-decomposed audit, delegated scoring, single JSON blueprint via PR), when to use it (explicit quoted trigger phrases), and how it differs from its siblings. Minor verbosity from its length, but no fluff or over-claims.

DimensionReasoningScore

Specificity

Lists multiple concrete actions — finds every model invocation (SDK/API call, sub-agent dispatch, agent frontmatter), decomposes by intent, delegates each to model-right-sizer-dryrun, and merges into one schema-conformant JSON blueprint committed via PR — comprehensive and specific rather than vague.

5 / 5

Completeness

Explicitly answers both 'what' (per-call-site intent-decomposed audit delegated to the dry-run skill, merged into one JSON blueprint committed via PR) and 'when' (a dedicated 'Use when someone says...' clause with concrete trigger phrases).

5 / 5

Trigger Term Quality

Quotes several natural trigger phrases a user would actually say ('audit this repo's model calls', 'right-size the models in <repo> per call site', 'find every LLM call and right-size it', 'commit a model right-sizing blueprint for <repo>'), with synonym variation.

5 / 5

Distinctiveness Conflict Risk

Carves a clear niche and explicitly contrasts itself against the sibling skills model-right-sizer-dryrun and model-right-sizer-install, minimizing the chance of triggering for the wrong skill.

5 / 5

Total

20

/

20

Passed

Validation

75%

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

Validation12 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

skill_md_line_count

SKILL.md is long (850 lines); consider splitting into references/ and linking

Warning

frontmatter_unknown_keys

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

Warning

relative_links

Relative link issues: 6 suspicious

Warning

referenced_paths_exist

Referenced path issues: 1 missing

Warning

Total

12

/

16

Passed

Repository
Cloudzero/cloudzero-claude-marketplace
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.