jbvc/continuous-learning-v2
Instinct-based learning system that observes sessions via hooks, creates atomic instincts with confidence scoring, and evolves them into skills/commands/agents.
40
40%
Does it follow best practices?
Impact
Pending
No eval scenarios have been run
Advisory
Suggest reviewing before use
Quality
Discovery
17%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 relies heavily on internal jargon and abstract concepts that would be opaque to both users and Claude when selecting among skills. It lacks a 'Use when...' clause and contains no natural trigger terms that a user would employ. The conceptual framework described is too abstract to clearly communicate concrete capabilities.
Suggestions
Add an explicit 'Use when...' clause describing concrete scenarios, e.g., 'Use when the user wants to automatically learn patterns from past sessions and generate reusable skills or commands.'
Replace jargon like 'atomic instincts' and 'instinct-based learning system' with plain-language descriptions of what the skill actually does, e.g., 'Automatically observes coding sessions, identifies recurring patterns, and generates reusable skills, commands, or agents with confidence scores.'
Include natural trigger terms users might say, such as 'learn from sessions', 'auto-generate skills', 'pattern detection', 'session observation', or 'evolve workflows'.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | It names some actions ('observes sessions via hooks', 'creates atomic instincts with confidence scoring', 'evolves them into skills/commands/agents') but these are abstract/conceptual rather than concrete user-facing actions. Terms like 'instinct-based learning system' and 'atomic instincts' are jargon rather than specific capabilities. | 2 / 3 |
Completeness | While there is a vague 'what' (observes sessions, creates instincts, evolves them), there is no 'when' clause or any explicit trigger guidance for when Claude should select this skill. The missing 'Use when...' clause caps this at 2 per the rubric, and the 'what' is also weak, so it scores 1. | 1 / 3 |
Trigger Term Quality | The description uses highly specialized jargon ('instinct-based learning', 'atomic instincts', 'confidence scoring', 'hooks') that users would almost never naturally say. There are no natural trigger terms a user would use when needing this functionality. | 1 / 3 |
Distinctiveness Conflict Risk | The highly specialized terminology ('instinct-based learning', 'atomic instincts') makes it somewhat distinctive from other skills, but the vague scope ('skills/commands/agents') could overlap with many meta-programming or automation skills. | 2 / 3 |
Total | 6 / 12 Passed |
Implementation
35%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This skill is overly verbose, spending significant tokens on marketing comparisons (v1 vs v2), architectural philosophy, and explanatory content that doesn't help Claude execute tasks. While it provides some concrete configuration examples, the core functionality (pattern detection, instinct creation, evolution) remains abstractly described rather than concretely implemented. The document would benefit greatly from being split into a lean overview with references to detailed docs.
Suggestions
Cut the v1 vs v2 comparison table, 'Why Hooks vs Skills' section, backward compatibility, privacy section, and tagline — these don't help Claude execute the skill. This alone would halve the document.
Add validation steps: how to verify hooks are firing (e.g., check observations.jsonl has entries), how to confirm the observer agent is running, and error recovery if instincts aren't being created.
Move configuration reference, file structure, and confidence scoring details into separate reference files (e.g., CONFIG.md, REFERENCE.md) and link to them from a concise overview.
Provide executable code or concrete implementation for the core operations (pattern detection, instinct creation, evolution clustering) rather than describing them with flow diagrams.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Extremely verbose at ~250+ lines. Includes unnecessary marketing-style content (tagline, comparison table with v1, 'Why Hooks vs Skills' explanation section, backward compatibility section, privacy section). The 'What's New in v2' table and philosophical explanations about why hooks are better are not actionable — Claude doesn't need to be sold on the architecture. Much of this could be cut or moved to reference docs. | 1 / 3 |
Actionability | Provides concrete JSON configuration for hooks and bash commands for directory setup, which is good. However, the slash commands (/instinct-status, /evolve) are referenced without showing their implementation. The instinct YAML format is shown but there's no executable code for the observer agent, pattern detection, or evolution clustering — the core functionality is described abstractly rather than implemented. | 2 / 3 |
Workflow Clarity | The Quick Start provides a 3-step sequence (enable hooks → initialize directories → use commands), and the ASCII flow diagram shows the pipeline. However, there are no validation checkpoints — no way to verify hooks are firing correctly, no way to confirm observations are being recorded, no error recovery steps if the observer agent fails. For a system involving background agents and file manipulation, this is a significant gap. | 2 / 3 |
Progressive Disclosure | References external links (Skill Creator app, GitHub repo, Twitter thread) but doesn't split its own verbose content into separate files. The configuration, file structure, confidence scoring details, and backward compatibility sections could all be in separate reference files. The main SKILL.md tries to be both overview and complete reference, resulting in a monolithic document. | 2 / 3 |
Total | 7 / 12 Passed |
Validation
90%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 10 / 11 Passed
Validation for skill structure
| Criteria | Description | Result |
|---|---|---|
frontmatter_unknown_keys | Unknown frontmatter key(s) found; consider removing or moving to metadata | Warning |
Total | 10 / 11 Passed | |
Reviewed
Table of Contents