do-in-steps

Execute one complex task as ordered, dependent steps run sequentially, passing context from each step to the next, with per-step LLM-as-a-judge verification. Use when later steps depend on the results of earlier ones.

Quality

85%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Critical

Do not install without reviewing

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.

Highly actionable and clearly sequenced with strong validation feedback loops, but the 1500-line monolithic body repeats dispatch flow and prompt templates across sections and examples and uses no bundle files for progressive disclosure.

Suggestions

Move the three full worked examples and the repeated prompt templates into reference files (e.g. references/examples.md, references/prompts.md) and link to them from the main body to cut duplication and token cost.

Consolidate the meta-judge/judge dispatch instructions so the flow is specified once (§3.3–§3.5 currently restate it); reference that single source from the examples instead of re-printing it.

Trim the verbatim re-printing of judge prompts inside Example 1 and Example 2 — a short summary of what was sent would carry the same point at a fraction of the length.

Dimension	Reasoning	Score
Conciseness	The body is mostly operational detail Claude does not already know, but at 1500 lines it repeats the same meta-judge/judge dispatch flow and full prompt templates across §3.3–§3.5 and again inside three complete examples, which is noticeable padding that could be tightened.	3 / 5
Actionability	Provides copy-paste-ready prompt templates, exact Task-tool dispatch structure with tool-call ordering, report naming conventions, and concrete score-threshold decision trees covering the common sequential-execution cases.	5 / 5
Workflow Clarity	A clearly sequenced Phase 1–4 process with explicit validation checkpoints (judge verification, score thresholds, retry loops, escalation ladder), error-recovery feedback loops, and checklists (RED FLAGS / ALWAYS).	5 / 5
Progressive Disclosure	Section headers give it structure, but it is a monolithic 1500-line SKILL.md with no bundle files and no references to separate files; large repeated prompt templates and three full examples are inlined rather than split out.	3 / 5
	Total	16 / 20 Passed

Description

92%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 voice, concrete actions, and an explicit Use-when trigger that distinguishes it from generic orchestration skills. Only minor synonym/keyword breadth is missing for a process-oriented skill.

Dimension	Reasoning	Score
Specificity	Names multiple concrete actions — "ordered, dependent steps run sequentially", "passing context from each step to the next", "per-step LLM-as-a-judge verification" — giving comprehensive coverage of the skill's mechanism.	5 / 5
Completeness	Explicitly answers both what (execute a complex task as ordered dependent steps with context passing and judge verification) and when ("Use when later steps depend on the results of earlier ones") with a concrete trigger phrase.	5 / 5
Trigger Term Quality	The "Use when later steps depend on the results of earlier ones" clause is a natural phrase users would say, but synonym coverage is limited for a process skill (no variants like "multi-step", "pipeline", "chain").	4 / 5
Distinctiveness Conflict Risk	The sequential-dependent-steps-with-judge-verification niche is clearly staked out and unlikely to fire for unrelated skills; the trigger is specific to dependent multi-step work.	5 / 5
	Total	19 / 20 Passed

Validation

87%

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 — 14 / 16 Passed

Validation for skill structure

Criteria	Description	Result
skill_md_line_count	SKILL.md is long (1503 lines); consider splitting into references/ and linking	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	14 / 16 Passed

Repository: NeoLabHQ/context-engineering-kit
Path: plugins/sadd/skills/do-in-steps/SKILL.md
Commit: 8539779

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.