create-agent

Create a new Hindsight-powered subagent with long-term memory. Use when the user wants a specialized agent that learns and remembers across sessions.

Quality

95%

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

100%Weight 40%Scale 1-3

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

A tight, highly actionable skill body with concrete tool calls, a complete subagent template, clear sequenced workflows including an error-recovery loop, and well-organized self-contained sections needing no external references.

Dimension	Reasoning	Score
Conciseness	The body is lean and procedural — terse tool signatures, a compact template, and short rules — with the only prose being necessary Hindsight-specific guidance; it assumes Claude's competence and does not pad with concepts Claude already knows.	3 / 3
Actionability	It gives concrete, copy-paste-ready guidance: exact tool calls with arguments (agent_knowledge_create_page(page_id, name, source_query)), a full subagent file template, and a precise output path (~/.claude/agents/<name>.md).	3 / 3
Workflow Clarity	Both invocation modes are clearly sequenced with numbered steps, plus an explicit error-recovery feedback loop (large page → Read the saved path → parse or skip) and verification steps after creation ('Confirm the subagent file was written').	3 / 3
Progressive Disclosure	No bundle files are needed; the content is self-contained and organized into clear sections (invocation modes, template, rules, after creation), with the inline template being the core deliverable that correctly stays in SKILL.md.	3 / 3
	Total	12 / 12 Passed

Description

90%Weight 40%Scale 1-3

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 concise, well-structured description that clearly states both capability and trigger with natural language and a distinct niche. Its only gap is specificity — it names a single overarching action rather than enumerating the concrete operations the skill performs.

Dimension	Reasoning	Score
Specificity	It names the domain ('Hindsight-powered subagent with long-term memory') and one concrete action ('Create a new ... subagent'), but does not list multiple specific actions like ingesting files, creating knowledge pages, or binding a bank, so it stops short of the 'multiple concrete actions' anchor.	2 / 3
Completeness	It answers both 'what' ('Create a new Hindsight-powered subagent with long-term memory') and 'when' ('Use when the user wants a specialized agent that learns and remembers across sessions') with an explicit trigger, matching the top anchor.	3 / 3
Trigger Term Quality	The trigger clause uses natural phrases a user would actually say — 'specialized agent that learns and remembers across sessions' — giving good coverage of natural terms rather than only jargon.	3 / 3
Distinctiveness Conflict Risk	The 'Hindsight-powered subagent with long-term memory' niche is distinct and unlikely to fire for unrelated skills, matching the clear-niche anchor.	3 / 3
	Total	11 / 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
allowed_tools_field	'allowed-tools' contains unusual tool name(s)	Warning

	Total	15 / 16 Passed

Repository: vectorize-io/hindsight
Path: hindsight-integrations/claude-code/skills/create-agent/SKILL.md
Commit: ed120a2

Reviewed: 7 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.