erne-learn

Data Community

Repo stars 36
License MIT
Author repo everything-react-native-expo

Interpretation is structured for decision-making; original keeps the upstream SKILL.md unchanged.

Decide Fit First

Core job: ERNE — Manual skill generation — runs continuous-learning-v2 scripts to extract patterns from session You are executing the /ern…
Best fit: Use it when the task has reusable inputs, steps, and validation criteria rather than a one-off answer.
Avoid forcing it: If the source lacks commands, platform support, or external-service evidence, keep those fields unknown instead of guessing.

Design Intent

Structure: The skill is organized around “What This Does”, “Process”, “When to Use”, “Note”, showing how the author expects the agent to judge fit, collect context, and produce verifiable output.
Trigger evidence: Prioritize the author’s wording around when to use it, what context to collect, and what output shape to produce.
Evidence boundary: Author text states facts, repository files prove commands and paths, and Fluxly only adds fit, limits, and usage judgment.

How To Use It

Inputs: Provide target material, scope, expected result, forbidden changes, and validation method.
Invocation: Name erne-learn directly; if the source includes slash commands, start with the command and then add task context.
Validation: Start small and check whether the result follows “What This Does / Process / When to Use” before expanding.

Boundaries And Review

Dependencies: It usually needs no extra API key, so start with a small validation task.
Permissions: Declared permissions include read / write / shell-exec; ask the agent to state file, command, and rollback boundaries before acting.
Quality bar: A useful result names the deliverable, evidence, and next action. Generic prose means the task needs tighter context.

/erne-learn — Generate Skills from Session

You are executing the /erne-learn command. This is script-driven — it runs the continuous-learning-v2 pipeline directly.

What This Does

Analyzes the current coding session to extract reusable patterns and save them as skills or rule amendments for future sessions. This is the manual trigger for what PostToolUse hooks do automatically.

Process

Run the extraction script:

node .claude/hooks/scripts/extract-session-patterns.js

This script:

Scans recent tool calls and file changes
Identifies patterns (repeated fixes, style corrections, common architectures)
Compares against existing rules and skills
Generates candidates for new content

Review candidates: The script outputs a list of potential learnings:

[PATTERN] Zustand store always uses immer middleware → Suggest rule amendment
[PATTERN] All screens use SafeAreaView wrapper → Suggest coding-style rule
[PATTERN] API calls always retry 3 times → Suggest pattern rule
[SKILL] Complex form validation flow → Suggest skill creation

Approve or reject each candidate: For each candidate, decide:

Approve → Script writes to .claude/rules/ or .claude/skills/
Reject → Skip this pattern
Edit → Modify before saving

Validate new content:

node .claude/hooks/scripts/validate-content.js

Ensures new rules/skills have valid frontmatter and don't conflict with existing content.

When to Use

After a long coding session where you established new patterns
When you notice yourself repeatedly making the same corrections
After integrating a new library and establishing conventions
Periodically to capture accumulated project knowledge

Note

The automatic PostToolUse hook (continuous-learning-v2.cjs) does lightweight extraction after every tool call. This /erne-learn command runs a comprehensive analysis that catches patterns the real-time hook might miss.

Fluxly profile Author and license come from source; runtime, permissions, and network are Fluxly detections or estimates

Fluxly category: Data
Author-declared agents: No explicit declaration found; this is not inferred or tested compatibility
Static check: 94 / 100 · heuristic scan, not runtime safety proof
Author / version / license: @JubaKitiashvili · MIT
Fluxly token estimate: Lean
Fluxly setup estimate: Guided setup
External API key: No requirement detected
Detected OS requirements: macOS · Linux · Windows
Runtime requirements: Node.js
Detected file/system behavior: Read-only
Write / modify
Shell exec
Detected network behavior: Local-only
Install commands: None (reference only)

Profile is derived at build time from SKILL.md and install vectors. Subject to drift from author intent.

Heads up: 未限定 allowed-tools，默认拥有全部工具权限。

Output preview erne-learn.preview

The current SKILL.md does not define a fixed output example.

SKILL.md SKILL.md

# /erne-learn — Generate Skills from Session

You are executing the `/erne-learn` command. This is **script-driven** — it runs the continuous-learning-v2 pipeline directly.

## What This Does

Analyzes the current coding session to extract reusable patterns and save them as skills or rule amendments for future sessions. This is the manual trigger for what `PostToolUse` hooks do automatically.

## Process

1. **Run the extraction script:**

```bash
node .claude/hooks/scripts/extract-session-patterns.js
```

This script:
- Scans recent tool calls and file changes
- Identifies patterns (repeated fixes, style corrections, common architectures)
- Compares against existing rules and skills
- Generates candidates for new content

2. **Review candidates:**
The script outputs a list of potential learnings:
```
[PATTERN] Zustand store always uses immer middleware → Suggest rule amendment
[PATTERN] All screens use SafeAreaView wrapper → Suggest coding-style rule
[PATTERN] API calls always retry 3 times → Suggest pattern rule
[SKILL] Complex form validation flow → Suggest skill creation
```

3. **Approve or reject each candidate:**
For each candidate, decide:
- **Approve** → Script writes to `.claude/rules/` or `.claude/skills/`
- **Reject** → Skip this pattern
- **Edit** → Modify before saving

4. **Validate new content:**
```bash
node .claude/hooks/scripts/validate-content.js
```
Ensures new rules/skills have valid frontmatter and don't conflict with existing content.

## When to Use
- After a long coding session where you established new patterns
- When you notice yourself repeatedly making the same corrections
- After integrating a new library and establishing conventions
- Periodically to capture accumulated project knowledge

## Note

…

Evidence boundary and execution chain

Author text anchors workflow facts; Fluxly only indexes current sections, terms, files, and commands.

sections -> What This Does → Process → When to Use → Note
terms    -> script-driven · Run the extraction script · Review candidates · Approve or reject each candidate · Approve · Reject · Edit · Validate new content
files/cmd -> /erne-learn · PostToolUse · .claude/rules/ · .claude/skills/ · continuous-learning-v2.cjs
body sha256 -> cac46f7a5f00