aif-loop

Run a strict multi-iteration Reflex Loop with phases (PLAN, PRODUCE||PREPARE, EVALUATE, CRITIQUE, REFINE) to improve an artifact until quality gates pass or iteration limits are reached. Use when user asks for iterative refinement, quality-gated generation, or "generate -> critique -> refine" loops.

Quality

84%

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

Quality

Content

85%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.

A highly actionable, well-structured orchestration skill with concrete JSON examples, explicit phase contracts, and clean progressive disclosure into verified reference files. Its only meaningful weakness is conciseness — some precedence-contract repetition and explanatory prose could be trimmed without losing clarity.

Suggestions

Consolidate the stop-condition precedence order into a single canonical statement; the cross-file sync discussion ('repeated verbatim in docs/loop.md and subagents/...') is meta-commentary that adds tokens without aiding execution.

Trim rhetorical flourishes like 'by definition rather than by accident' and 'in the same breath' to plain statements of the invariant.

Move the budget's inline restatement (Step 5) to a one-line pointer to references/ACTIVE-TIME-BUDGET.md, since the full contract already lives there.

Dimension	Reasoning	Score
Conciseness	Mostly efficient for a complex orchestration skill, but includes repeated emphasis (the precedence order restated and discussed across multiple files) and rhetorical flourishes ('by definition rather than by accident', 'in the same breath') that could be tightened.	3 / 5
Actionability	Fully executable guidance: concrete run.json/current.json JSON examples, exact persistence file paths, enumerated event names, command-mode specs, and phase-output formats that are copy-paste ready.	5 / 5
Workflow Clarity	Nine clearly sequenced steps with explicit validation checkpoints (phase_error retry-once, corrupted run.json recovery, skill-context enforcement) and a feedback loop for destructive `clean` requiring user confirmation.	5 / 5
Progressive Disclosure	Clear overview body with well-signaled, one-level-deep references to real files (ACTIVE-TIME-BUDGET.md, CRITERIA-TEMPLATES.md, PHASE-CONTRACTS.md, RULE-SCHEMA.md, TERMINAL-REPORT.md, CONTEXT-MANAGEMENT.md); detailed contracts are appropriately split out.	5 / 5
	Total	18 / 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 that clearly states both capability and trigger conditions in third-person imperative voice with concrete phase names and natural trigger phrases. The main weakness is the generic 'artifact' domain, which leaves a small gap in specificity and distinctiveness.

Dimension	Reasoning	Score
Specificity	Names several concrete actions — the six phases (PLAN, PRODUCE, PREPARE, EVALUATE, CRITIQUE, REFINE) and 'improve an artifact until quality gates pass' — but the artifact domain itself stays generic, leaving minor coverage gaps.	4 / 5
Completeness	Explicitly answers both what (run a multi-iteration Reflex Loop through six phases to improve an artifact) and when (a clear 'Use when...' clause with concrete trigger phrases).	5 / 5
Trigger Term Quality	Includes natural phrases users would say ('iterative refinement', 'quality-gated generation', 'generate -> critique -> refine'), with good keyword coverage, though a few common synonyms are missing.	4 / 5
Distinctiveness Conflict Risk	The 'Reflex Loop' framing and 'generate -> critique -> refine' trigger carve a fairly distinct niche, though 'quality-gated generation' has minor overlap with broader generation skills.	4 / 5
	Total	17 / 20 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 — 13 / 16 Passed

Validation for skill structure

Criteria	Description	Result
skill_md_line_count	SKILL.md is long (507 lines); consider splitting into references/ and linking	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning
referenced_paths_exist	Referenced path issues: 2 missing	Warning

	Total	13 / 16 Passed

Repository: lee-to/ai-factory
Path: skills/aif-loop/SKILL.md
Commit: e144f95

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.