Tessl • The package manager for agent skills

Blog Docs Log inGet started

SkillPublic

external-model-delegation

UserPromptSubmit hook pattern that classifies prompts into six cost tiers and delegates to external model CLIs

Quality

65%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./skills/external-model-delegation/SKILL.md

Quality

Content

68%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 efficient and actionable, with concrete commands and a usable template, and is well-structured for inline reference. Its main gaps are a happy-path-only workflow lacking explicit validation/fallback steps and time-sensitive pricing mixed into the main content.

Suggestions

Add an explicit validation/fallback step to the routing flow, e.g. "If the delegation script exits non-zero, surface the stderr error and fall back to handling natively (CLAUDE tier)."

Move the per-M-token pricing into a clearly labeled, easily-updatable section (or a separate reference file) so volatile cost data doesn't clutter the core routing guidance.

Optionally extract the per-script implementation details (qwen3 Ollama curl, deepseek httpx, kimi exec) into a reference file to bring SKILL.md under the simple-skill threshold and lift progressive disclosure.

Dimension	Reasoning	Score
Conciseness	The body is lean, assumes Claude's competence, and uses tables and a compact bash template rather than explaining known concepts. Not a 5 because the pricing table (specific per-M-token dollar amounts) is time-sensitive data not placed in a deprecated/old-patterns section.	4 / 5
Actionability	Exact delegation commands (e.g. `qwen3 "prompt"`, `deepseek --flash "prompt"`), a copy-paste bash template, a numbered script contract, and concrete env vars make it mostly executable. Not a 5 because the real qwen3/kimi/deepseek scripts are only described in one-line comments rather than shown as complete code.	4 / 5
Workflow Clarity	The routing hook flow gives a clear 6-step sequence and the contract specifies exit-code/stderr error signaling, but the flow shows the happy path only with no explicit validation checkpoint or fallback for misclassification or script failure. Not a 4 because checkpoints are implicit rather than present.	3 / 5
Progressive Disclosure	Content is well-organized under clear section headers with tables and code blocks, and there are no broken or nested references. Not a 5 because at ~85 lines (above the 50-line simple-skill threshold) everything is inlined with no content split into reference files.	4 / 5
	Total	15 / 20 Passed

Description

46%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 states a clear, fairly specific purpose but relies on technical jargon and omits any explicit "when to use" guidance, capping completeness. It is distinctive in niche but weak on natural trigger terms.

Suggestions

Add a "Use when..." clause to the description (or ensure trigger guidance is visible) so it explicitly answers when to invoke the skill, e.g. "Use when routing prompts to cheaper external models via a UserPromptSubmit hook."

Replace pure jargon ("UserPromptSubmit hook pattern", "cost tiers") with at least one natural user phrase such as "delegate prompts to external models" to improve trigger-term quality.

Add a concrete action or two (e.g. "injects delegation context and returns the external model's output") to lift specificity toward comprehensive coverage.

Dimension	Reasoning	Score
Specificity	Names the domain ("UserPromptSubmit hook pattern") and two concrete actions ("classifies prompts into six cost tiers" and "delegates to external model CLIs"), but coverage is not comprehensive. Not a 4 because only two actions are listed with no detail on context injection or script spawning.	3 / 5
Completeness	It clearly answers "what" (classifies prompts into tiers and delegates to external CLIs) but the description field contains no "Use when..." clause, so per the rubric guideline completeness is capped at 3.	3 / 5
Trigger Term Quality	Keywords are largely technical jargon ("UserPromptSubmit hook", "cost tiers", "external model CLIs") rather than natural phrases a user would say; only "delegates to external model" is semi-natural. Not a 1 because there is one roughly user-facing phrase, but common synonyms are missing.	2 / 5
Distinctiveness Conflict Risk	It targets a clear niche (prompt-based delegation to external model CLIs via a hook) with low overlap risk. Not a 5 because the description lacks distinct explicit trigger phrases that would fully separate it from adjacent routing skills.	4 / 5
	Total	12 / 20 Passed

Validation

93%

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 — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	15 / 16 Passed

Repository: alinaqi/maggy
Path: skills/external-model-delegation/SKILL.md
Revision: 55aabe2

Reviewed: 20 days ago

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.

external-model-delegation

Table of Contents

Is this your skill?