loop-design-check

Design a goal-oriented agent loop, and review it for the ways loops go wrong — spinning and burning tokens, Goodhart-gaming the verifier, or running a wrong answer to completion. Two actions: (1) WRITE a loop — gate whether to build it, define a machine-decidable goal, pick the loop type, pick a skeleton; (2) REVIEW a loop — run it past five failure modes plus decidability, boundaries, fallback, judge independence, and keep-judgment-with-the-human red lines. Use when designing an autonomous agent loop, or when you already have one and worry it will spin, cheat, or run a wrong answer to the end. Complements the mechanism-layer loop skills (autonomous-loops, continuous-agent-loop) by covering the judgment layer they don't. 中文触发：写 loop、设计 loop、做一个 loop、检查 loop 对不对、loop 体检、loop 会不会跑飞、可判定目标、五个崩法、plan build judge。English triggers: design an agent loop, write a loop, check a loop, loop review, prevent a runaway loop, goal-oriented loop, decidable goal, plan/build/judge.

Quality

96%

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

Quality

Content

92%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 dense, well-structured design+review playbook with concrete frameworks, explicit validation checkpoints, and strong feedback loops. Its only weakness is mild editorial padding that slightly undercuts token efficiency. No bundle files are present, but the single-file structure is appropriate at this length.

Suggestions

Trim rhetorical aphorisms (e.g., 'the Ralph-Wiggum loop', the closing 'A loop only rewards someone who has already thought it through...') to recover tokens without losing actionable content.

The 'Lineage' footnote and repeated mechanism-layer carve-off (stated in both When-to-use and again at the end) could be consolidated into one mention.

Dimension	Reasoning	Score
Conciseness	Assumes Claude's intelligence (no basic-concept explaining) and is mostly efficient, but carries some preachy editorializing ('the Ralph-Wiggum loop', the closing aphorism) and minor rhetorical repetition that could be trimmed.	4 / 5
Actionability	Highly concrete and specific: a 4-condition build gate, a five-point goal framework, a servo/regulator loop-type table, a plan/build/judge role table, and a five-row review checklist each with explicit antibodies. Guidance is directly executable for a design skill.	5 / 5
Workflow Clarity	Both actions are clearly sequenced (Write = 5 numbered steps; Review = checklist) with explicit validation checkpoints and feedback loops: 'exit code is final', state stops at 'awaiting verification', 'Three failed retries -> escalate to a human', and 'Validate -> if errors fix and re-validate -> only when valid proceed'.	5 / 5
Progressive Disclosure	No bundle files exist; the entire skill is a single ~135-line SKILL.md with clear section headers, tables, and a worked example. At this size and organization the simple-skill exception applies and structure is clean with no nested references.	5 / 5
	Total	19 / 20 Passed

Description

100%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 is comprehensive, concrete, and explicit about both capability and trigger conditions, with bilingual trigger terms. It clearly distinguishes itself from sibling mechanism-layer loop skills. One minor cost is sheer length, but the density is justified by the specificity it carries.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions across both WRITE (gate, define machine-decidable goal, pick loop type, pick skeleton) and REVIEW (five failure modes, decidability, boundaries, fallback, judge independence, red lines), giving comprehensive coverage.	5 / 5
Completeness	Explicitly answers both 'what' (design and review agent loops; two actions enumerated) and 'when' ('Use when designing an autonomous agent loop, or when you already have one and worry it will spin, cheat, or run a wrong answer to the end') with concrete trigger phrases.	5 / 5
Trigger Term Quality	Comprehensive natural trigger coverage in both English ('design an agent loop, write a loop, check a loop, loop review, prevent a runaway loop, goal-oriented loop, decidable goal, plan/build/judge') and Chinese, including synonyms and phrases a user would actually say.	5 / 5
Distinctiveness Conflict Risk	Clear niche (the judgment layer of agent loops) and explicitly carves off the mechanism layer ('Complements the mechanism-layer loop skills...by covering the judgment layer they don't'), giving minimal conflict risk.	5 / 5
	Total	20 / 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
metadata_version	'metadata.version' is missing	Warning

	Total	15 / 16 Passed

Repository: affaan-m/ECC
Path: skills/loop-design-check/SKILL.md
Commit: d8409a4

Reviewed: 1 day 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.