oma-search

Intent-based search router with trust scoring. Routes queries to optimal channels (Context7 docs, native web search, gh/glab code search, Serena local) and attaches domain trust labels. Use for search, find, lookup, reference, docs, code search, and web research.

Quality

75%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./benchmarks/runs/oma/.agents/skills/oma-search/SKILL.md

Quality

Content

67%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 well structured for routing decisions — a clear route table, guardrails, fallback transitions, and concrete CLI commands — but it is padded with meta-framework scaffolding (SSL primitive mappings, resource-scope tables) and redundant restatements of the same routes. Progressive disclosure is well signaled on paper, though none of the referenced resource files are present in the bundle to back it.

Suggestions

Remove the 'Actions | SSL primitive' and 'Resource scope' tables (or compress them to a single line each): they add token cost without changing what Claude does, addressing the conciseness score.

Deduplicate the route information: the Routes table already carries tool, fallback, and trigger; drop the repeated restatements in 'Tools and instruments', 'Transitions', and guardrails 5-6, and collapse the References section's doubled prose-plus-bullet listing into one list.

Add one inline concrete example each for intent classification and trust scoring (e.g. a sample query -> route decision, and a domain -> trust label mapping), since the files that supposedly contain those procedures are not part of the bundle.

Dimension	Reasoning	Score
Conciseness	The body carries notable scaffolding overhead: the 'Actions \| SSL primitive' table (READ/SELECT/CALL_TOOL/INFER/NOTIFY) and the 'Resource scope' table (NETWORK/CODEBASE/PROCESS/MEMORY) add labels rather than actionable guidance, the routes are described three times (Routes table, Guardrails, Tools and instruments), and the References section lists every resource file twice (once in prose, once as bullets). Substantive routing content keeps it above 2, but it could clearly be tightened.	3 / 5
Actionability	Concrete, executable commands are present ('gh search code "<query>"', 'glab api "/search?scope=blobs&search=<query>"'), the docs route names the actual tool sequence ('resolve-library-id -> query-docs'), and invocation examples are given. Gaps remain: the web route says only 'runtime native search' and trust scoring has no inline procedure, so it is not fully copy-paste ready.	4 / 5
Workflow Clarity	The Default Workflow (Parse -> Route -> Collect -> Score -> Present) and scene sequence are clear, with an explicit VERIFY stage and failure/recovery transitions (docs falls back to web, web escalates to fetch strategies). Validation is stated abstractly ('Apply trust scoring and route-specific quality checks') and the concrete checklist lives in an external file, so explicit inline checkpoints are incomplete for a 5.	4 / 5
Progressive Disclosure	The References section signals each file with its purpose ('Execution steps', 'Intent classification', 'Trust registry', etc.) at one level of depth. It is held below 5 because the listing is duplicated (prose plus bullet list), and the referenced files (resources/.md, ../_shared/core/.md) do not exist in the provided bundle, so navigation cannot be verified and some paths escape the skill directory.	4 / 5
	Total	15 / 20 Passed

Description

83%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.

A strong description: third-person, concise, concrete about channels and trust labels, with an explicit trigger clause. Its main residual risk is that very broad trigger verbs ('search', 'find', 'lookup') make it fire on almost any information-seeking request, overlapping native search capabilities.

Dimension	Reasoning	Score
Specificity	The description lists several concrete actions ('Routes queries to optimal channels (Context7 docs, native web search, gh/glab code search, Serena local) and attaches domain trust labels') with named channels. Falls short of 5 because the output side (ranked results, fallbacks) is left unspecified; above 3 because more than 1-2 concrete actions are named.	4 / 5
Completeness	Both 'what' (routes queries to channels and attaches domain trust labels) and 'when' (an explicit 'Use for search, find, lookup, reference, docs, code search, and web research' clause with concrete trigger phrases) are clearly and explicitly stated. Not 4 because the 'when' clause does list concrete triggers rather than being merely present.	5 / 5
Trigger Term Quality	'Use for search, find, lookup, reference, docs, code search, and web research' covers the natural trigger verbs well. A few common variations are missing (e.g. 'look up', 'documentation', 'find examples', 'how do I'), so it does not reach comprehensive synonym coverage.	4 / 5
Distinctiveness Conflict Risk	The intent-based routing plus trust-scoring framing carves a distinct niche with named channels. Trigger verbs like 'search', 'find', and 'docs' are extremely common and overlap with native web-search and docs tooling, keeping it below a 5.	4 / 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: first-fluke/oh-my-agent
Path: benchmarks/runs/oma/.agents/skills/oma-search/SKILL.md
Commit: ce19702

Reviewed: 4 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.