Tessl • The package manager for agent skills

SkillPublic

embedding-strategies

Select and optimize embedding models for semantic search and RAG applications. Use when choosing embedding models, implementing chunking strategies, or optimizing embedding quality for specific domains.

Quality

79%

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 ./plugins/llm-application-dev/skills/embedding-strategies/SKILL.md

Quality

Content

65%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 a well-structured, token-efficient overview with an exemplary progressive-disclosure handoff to a real, one-level reference file. It is weaker on actionability and workflow clarity: no executable guidance in SKILL.md itself and no concrete chunking parameters or validation checkpoints for the pipeline it depicts.

Suggestions

Inline one minimal copy-paste example (e.g. a 5-line voyage-3-large embed call) in the body so the core action is executable without opening the reference.

Add concrete chunking parameters (typical size ranges, overlap ratios, boundary heuristics for prose vs. code) instead of the "[Overlap, Size]" placeholder.

Add a short validation checkpoint to the pipeline, such as checking token counts against the model's max tokens before embedding and spot-checking similarity scores afterward.

Dimension	Reasoning	Score
Conciseness	Mostly lean — a dense model-comparison table and terse do's/don'ts with no explanation of known concepts — but the ASCII pipeline diagram states the obvious sequence, and the date-stamped "(2026)" model table is time-sensitive data not placed in an old-patterns/deprecated section.	4 / 5
Actionability	The model table is concrete decision data, but the body contains no executable code or commands, and chunking — a headline capability — has zero concrete parameters (only the placeholder "[Overlap, Size]"); all executable content is deferred to references/details.md.	3 / 5
Workflow Clarity	The pipeline diagram gives a rough conceptual sequence (Chunking → Preprocessing → Embedding) but no operational steps, commands, or validation checkpoints — e.g. nothing about verifying embedding quality or detecting token-limit truncation, which the do's/don'ts themselves flag as a risk.	3 / 5
Progressive Disclosure	A ~65-line overview with a single well-signaled, one-level-deep reference ("Full template library and detailed worked examples live in references/details.md"), verified to exist and to contain exactly the promised templates and worked examples; nothing that belongs in the reference is inlined.	5 / 5
	Total	15 / 20 Passed

Description

82%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 cleanly answers both what the skill does and when to use it, in third person, with strong natural trigger terms like RAG, semantic search, and chunking. Its only weaknesses are modest action specificity (select/optimize are generic verbs) and a few missing synonyms such as vector search or retrieval.

Suggestions

Enumerate one or two more concrete actions in the what-clause (e.g. "reduce embedding dimensions and compare model benchmarks") to lift specificity beyond select/optimize.

Add missing natural synonyms to the when-clause, such as "building vector search" or "working with a vector database".

Dimension	Reasoning	Score
Specificity	Names the domain and two concrete actions ("Select and optimize embedding models"), but coverage is not comprehensive — actions are fewer and more generic than the 4-anchor's several specific actions, and "implementing chunking strategies" appears only in the when-clause.	3 / 5
Completeness	Explicitly answers both: "Select and optimize embedding models for semantic search and RAG applications" (what) and "Use when choosing embedding models, implementing chunking strategies, or optimizing embedding quality for specific domains" (when, with concrete trigger phrases).	5 / 5
Trigger Term Quality	Good natural keyword coverage ("semantic search", "RAG", "embedding models", "chunking strategies"), but a few natural synonyms users would say — e.g. "vector search", "vector database", "retrieval" — are missing.	4 / 5
Distinctiveness Conflict Risk	A clear niche (embedding model selection and chunking for RAG/semantic search) with distinct triggers; only a much broader RAG/retrieval skill could overlap, so conflict risk is minimal.	5 / 5
	Total	17 / 20 Passed

Validation

100%

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 — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: wshobson/agents
Path: plugins/llm-application-dev/skills/embedding-strategies/SKILL.md
Revision: 0874a27

Reviewed: 21 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.

embedding-strategies

Table of Contents

Is this your skill?