agent-orchestration-improve-agent

Systematic improvement of existing agents through performance analysis, prompt engineering, and continuous iteration.

Quality

56%

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

Fix and improve this skill with Tessl

tessl review fix ./skills/antigravity-agent-orchestration-improve-agent/SKILL.md

Quality

Content

62%

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

The skill lays out a strong, well-sequenced optimization workflow with real validation and rollback feedback loops, but it is over-long, leans on concept explanation and placeholder templates rather than executable guidance, and keeps everything inline instead of splitting detail into reference files.

Suggestions

Cut concept explanations Claude already knows (Constitutional AI principles, statistical-significance/Cohen's d primers, MAJOR/MINOR/PATCH semantics) down to the skill-specific decisions that matter.

Replace placeholder templates ('[X%]', '[User request]') with one concrete worked example using real tool calls and real expected output, and confirm the cited tools (context-manager, prompt-engineer, parallel-test-runner) are actually invokable.

Move the metric frameworks, test-category catalog, and version/rollback detail into separate reference files (e.g. references/metrics.md, references/testing.md) and keep SKILL.md as a concise overview with one-level-deep links.

Dimension	Reasoning	Score
Conciseness	The ~350-line body is structured and task-relevant, but it explains concepts Claude already knows (Constitutional AI, Cohen's d / power analysis, MAJOR/MINOR/PATCH versioning) and pads with long bullet enumerations — 'mostly efficient but includes some unnecessary explanation', the level-2 anchor.	2 / 3
Actionability	It gives structured guidance and named tool invocations ('Use: context-manager / Command: analyze-agent-performance $ARGUMENTS --days 30') and example templates, but the code blocks are placeholder-filled ('[X%]', '[User request]') rather than copy-paste executable, matching the level-2 'some concrete guidance but incomplete; pseudocode instead of executable code' anchor.	2 / 3
Workflow Clarity	Phases 1–4 are clearly sequenced with explicit validation checkpoints ('Validate with tests', staged rollout gates), explicit feedback loops (rollback triggers and a detect→alert→switch→analyze→re-test process), and checklists (Success Criteria, test categories), satisfying the level-3 anchor. It is not level 2 because validation is explicit, not implicit or missing.	3 / 3
Progressive Disclosure	No bundle files exist (references/scripts/assets are empty) and the entire ~350-line body is inline in one file, yet the content has decent phase/subsection organization — so it is the level-2 'some structure but content that should be separate is inline' rather than a level-1 poorly-organized monolith or a level-3 split-with-references skill.	2 / 3
	Total	9 / 12 Passed

Description

50%

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 states a clear purpose and domain but stops short of giving Claude an explicit trigger for when to use it, and its action list is high-level rather than concrete. It is competent but indistinct within a likely skill family.

Suggestions

Add an explicit 'Use when...' clause naming concrete triggers, e.g. 'Use when improving an existing agent's performance, diagnosing failure modes, or running A/B prompt tests — not for building agents from scratch.'

Replace abstract categories ('performance analysis', 'prompt engineering') with more concrete actions ('analyze failure traces, rewrite prompt instructions, run regression test suites').

Add natural-term variations users would actually say (e.g. 'agent not working', 'prompt quality', 'hallucinations', 'A/B test prompts') to strengthen trigger matching.

Dimension	Reasoning	Score
Specificity	Quotes 'Systematic improvement of existing agents through performance analysis, prompt engineering, and continuous iteration' — it names a domain plus three actions, but those actions ('performance analysis', 'prompt engineering', 'continuous iteration') are abstract categories rather than the concrete specific actions the level-3 anchor requires.	2 / 3
Completeness	It clearly answers 'what' (systematic improvement of existing agents via three techniques) but never answers 'when' — there is no 'Use when...' clause or equivalent explicit trigger guidance, which per the judging guidelines caps completeness at 2.	2 / 3
Trigger Term Quality	Terms like 'agent performance' and 'prompt engineering' are reasonably natural, but coverage is thin, there are no common variations, and the description lacks any trigger framing, matching the level-2 anchor of 'some relevant keywords but missing common variations'.	2 / 3
Distinctiveness Conflict Risk	'Improvement of existing agents' is a defined niche, but it sits inside an 'agent-orchestration' skill family and lacks distinct trigger language, so it could still overlap with neighboring skills — the level-2 'somewhat specific but could still overlap' anchor.	2 / 3
	Total	8 / 12 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
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	15 / 16 Passed

Repository: boisenoise/skills-collections
Path: skills/antigravity-agent-orchestration-improve-agent/SKILL.md
Commit: c27161a

Reviewed: about 3 hours 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.