do-in-parallel

Launch multiple sub-agents in parallel to execute tasks across files or targets with intelligent model selection, quality-focused prompting, and meta-judge → LLM-as-a-judge verification

Quality

36%

Does it follow best practices?

Impact

—

No eval scenarios have been run

Securityby

Passed

No known issues

Optimize this skill with Tessl

npx tessl skill review --optimize ./plugins/sadd/skills/do-in-parallel/SKILL.md

Quality

Content

55%

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

This skill demonstrates excellent actionability and workflow clarity with concrete, executable templates and a well-structured multi-phase process with proper validation checkpoints and error recovery. However, it is severely undermined by extreme verbosity — the three examples alone are massively redundant, repeating nearly identical prompt templates with minor variations. The monolithic structure with no progressive disclosure makes it a poor fit for a context window, consuming enormous token budget when most of the content is repetitive.

Suggestions

Extract the three full examples into a separate EXAMPLES.md file, keeping only a brief summary table in the main SKILL.md showing the grouping type, agent counts, and key differences between scenarios.

Consolidate the repeated prompt templates (meta-judge, judge, implementor) into a single TEMPLATES.md reference file, with the main SKILL.md showing just one canonical example of each template type.

Remove explanatory content Claude already knows — e.g., the paragraph explaining what parallel execution is, the definition of 'fresh context', and basic concepts like what PDF libraries or CI pipelines are. The <context> section could be reduced to 3-4 bullet points.

Eliminate redundancy in the ALWAYS/NEVER lists and Best Practices section, which repeat rules already stated in the process phases (e.g., 'dispatch meta-judges in parallel' appears at least 5 times in different sections).

Dimension	Reasoning	Score
Conciseness	This skill is extremely verbose at ~1000+ lines. It extensively repeats the same prompt templates across three full examples with minimal variation, explains concepts Claude already understands (what parallel execution is, what independence means), and includes massive amounts of redundant content. The three examples alone consume the majority of the document and could be replaced with one concise example plus a brief note on variations. The requirement grouping analysis, meta-judge templates, and judge templates are each stated multiple times in slightly different forms.	1 / 3
Actionability	The skill provides highly concrete, copy-paste-ready prompt templates for every agent type (meta-judge, implementor, judge), specific decision trees for model selection and requirement grouping, exact structured output formats, and detailed dispatch patterns. Every phase has executable templates with placeholder variables clearly marked.	3 / 3
Workflow Clarity	The multi-phase workflow is clearly sequenced (Parse → Analyze → Meta-Judge → Implement → Judge → Retry → Summarize) with explicit validation checkpoints at every stage. The retry logic includes max retry limits, failure isolation, shared group retry specifics, and clear escalation paths. The independence validation checklist and verdict parsing logic provide strong feedback loops for error recovery.	3 / 3
Progressive Disclosure	The entire skill is a monolithic wall of text with no references to external files despite being well over 1000 lines. The three full examples (each 100+ lines of near-identical prompt templates) should be in a separate EXAMPLES.md. The prompt templates could be in a TEMPLATES.md. The decision trees and tables for model selection could be in a REFERENCE.md. Everything is inline, making the document extremely difficult to navigate.	1 / 3
	Total	8 / 12 Passed

Description

17%

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 overly technical and architecture-focused, reading more like an internal system design note than a skill description meant to help Claude select the right tool. It lacks a 'Use when...' clause, natural user-facing trigger terms, and concrete examples of what tasks it actually performs. The buzzwords ('intelligent model selection', 'quality-focused prompting') add noise without clarity.

Suggestions

Add an explicit 'Use when...' clause with natural trigger terms, e.g., 'Use when the user asks to process multiple files in parallel, batch operations, or run the same task across many targets.'

Replace jargon like 'meta-judge → LLM-as-a-judge verification' with user-understandable language, e.g., 'automatically verifies output quality across parallel tasks.'

List concrete actions the skill performs, e.g., 'Runs code modifications, linting, or analysis across multiple files simultaneously, then validates results for correctness.'

Dimension	Reasoning	Score
Specificity	It names some actions like 'launch sub-agents in parallel', 'execute tasks across files or targets', 'model selection', and 'verification', but these are described at a high architectural level rather than listing concrete user-facing actions. Terms like 'intelligent model selection' and 'quality-focused prompting' are somewhat buzzwordy.	2 / 3
Completeness	The description partially addresses 'what' (launching parallel sub-agents with verification) but completely lacks any 'when' clause or explicit trigger guidance. There is no 'Use when...' or equivalent, which per the rubric should cap completeness at 2, and the 'what' itself is vague enough to warrant a 1.	1 / 3
Trigger Term Quality	The description uses technical jargon like 'sub-agents', 'meta-judge', 'LLM-as-a-judge verification' that users would rarely naturally say. It lacks common user-facing trigger terms — a user needing parallel task execution would more likely say 'run in parallel', 'batch process', or 'process multiple files'.	1 / 3
Distinctiveness Conflict Risk	The mention of parallel sub-agents and LLM-as-a-judge verification gives it some distinctiveness, but 'execute tasks across files or targets' is broad enough to overlap with many file-processing or task-orchestration skills.	2 / 3
	Total	6 / 12 Passed

Validation

81%

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 — 9 / 11 Passed

Validation for skill structure

Criteria	Description	Result
skill_md_line_count	SKILL.md is long (2209 lines); consider splitting into references/ and linking	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	9 / 11 Passed

Repository: NeoLabHQ/context-engineering-kit
Commit: 3711edf

Reviewed: 5 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.