skill-idea-miner

Data Community

Repo stars 1,568
Author repo claude-trading-skills

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

Decide Fit First

Core job: Mine Claude Code session logs for skill idea candidates. Use when running the weekly skill generation pipeline to extract, score…
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 “When to Use”, “Prerequisites”, “Workflow”, “Quick Start”, 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 skill-idea-miner 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 “When to Use / Prerequisites / Workflow” 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; 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.

Skill Idea Miner

Automatically extract skill idea candidates from Claude Code session logs, score them for novelty, feasibility, and trading value, and maintain a prioritized backlog for downstream skill generation.

When to Use

Weekly automated pipeline run (Saturday 06:00 via launchd)
Manual backlog refresh: python3 scripts/run_skill_generation_pipeline.py --mode weekly
Dry-run to preview candidates without LLM scoring

Prerequisites

Python 3.10+ with pyyaml package
Claude CLI installed and authenticated (claude --version to verify)
Session logs in ~/.claude/projects/<project>/ (created automatically by Claude Code)
No API keys required (uses Claude CLI for LLM calls)

Workflow

Quick Start

# Dry-run: preview mined candidates without LLM scoring
python3 scripts/mine_session_logs.py --dry-run --output-dir reports/

# Full mining with scoring (requires Claude CLI)
python3 scripts/mine_session_logs.py --output-dir reports/

# Score existing candidates
python3 scripts/score_ideas.py \
  --candidates reports/raw_candidates.yaml \
  --output-dir logs/

Stage 1: Session Log Mining

Enumerate session logs from allowlist projects in ~/.claude/projects/
Filter to past 7 days by file mtime, confirm with timestamp field
Extract user messages (type: "user", userType: "external")
Extract tool usage patterns from assistant messages
Run deterministic signal detection:
- Skill usage frequency (skills/*/ path references)
- Error patterns (non-zero exit codes, is_error flags, exception keywords)
- Repetitive tool sequences (3+ tools repeated 3+ times)
- Automation request keywords (English and Japanese)
- Unresolved requests (5+ minute gap after user message)
Invoke Claude CLI headless for idea abstraction
Output raw_candidates.yaml

Stage 2: Scoring and Deduplication

Load existing skills from skills/*/SKILL.md frontmatter
Deduplicate via Jaccard similarity (threshold > 0.5) against:
- Existing skill names and descriptions
- Existing backlog ideas
Score non-duplicate candidates with Claude CLI:
- Novelty (0-100): differentiation from existing skills
- Feasibility (0-100): technical implementability
- Trading Value (0-100): practical value for investors/traders
- Composite = 0.3 * Novelty + 0.3 * Feasibility + 0.4 * Trading Value
Merge scored candidates into logs/.skill_generation_backlog.yaml

Output Format

raw_candidates.yaml

generated_at_utc: "2026-03-08T06:00:00Z"
period: {from: "2026-03-01", to: "2026-03-07"}
projects_scanned: ["claude-trading-skills"]
sessions_scanned: 12
candidates:
  - id: "raw_2026w10_001"
    title: "Earnings Whispers Image Parser"
    source_project: "claude-trading-skills"
    evidence:
      user_requests: ["Extract earnings dates from screenshot"]
      pain_points: ["Manual image reading"]
      frequency: 3
    raw_description: "Parse Earnings Whispers screenshots to extract dates."
    category: "data-extraction"

Backlog (logs/.skill_generation_backlog.yaml)

updated_at_utc: "2026-03-08T06:15:00Z"
ideas:
  - id: "idea_2026w10_001"
    title: "Earnings Whispers Image Parser"
    description: "Skill that parses Earnings Whispers screenshots..."
    category: "data-extraction"
    scores: {novelty: 75, feasibility: 60, trading_value: 80, composite: 73}
    status: "pending"

Resources

references/idea_extraction_rubric.md — Signal detection criteria and scoring rubric
scripts/mine_session_logs.py — Session log parser
scripts/score_ideas.py — Scorer and deduplicator

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: 88 / 100 · heuristic scan, not runtime safety proof
Author / version / license: @tradermonty · no license declared
Fluxly token estimate: Lean
Fluxly setup estimate: Plug-and-play
External API key: No requirement detected
Detected OS requirements: Unspecified
Runtime requirements: Python >=3.10
Detected file/system behavior: Read-only
Write / modify
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 skill-idea-miner.preview

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

SKILL.md SKILL.md

# Skill Idea Miner

Automatically extract skill idea candidates from Claude Code session logs,
score them for novelty, feasibility, and trading value, and maintain a
prioritized backlog for downstream skill generation.

## When to Use

- Weekly automated pipeline run (Saturday 06:00 via launchd)
- Manual backlog refresh: `python3 scripts/run_skill_generation_pipeline.py --mode weekly`
- Dry-run to preview candidates without LLM scoring

## Prerequisites

- **Python 3.10+** with `pyyaml` package
- **Claude CLI** installed and authenticated (`claude --version` to verify)
- **Session logs** in `~/.claude/projects/<project>/` (created automatically by Claude Code)
- No API keys required (uses Claude CLI for LLM calls)

## Workflow

### Quick Start

```bash
# Dry-run: preview mined candidates without LLM scoring
python3 scripts/mine_session_logs.py --dry-run --output-dir reports/

# Full mining with scoring (requires Claude CLI)
python3 scripts/mine_session_logs.py --output-dir reports/

# Score existing candidates
python3 scripts/score_ideas.py \
  --candidates reports/raw_candidates.yaml \
  --output-dir logs/
```

### Stage 1: Session Log Mining

1. Enumerate session logs from allowlist projects in `~/.claude/projects/`
2. Filter to past 7 days by file mtime, confirm with `timestamp` field
3. Extract user messages (`type: "user"`, `userType: "external"`)
4. Extract tool usage patterns from assistant messages
5. Run deterministic signal detection:
   - Skill usage frequency (`skills/*/` path references)
   - Error patterns (non-zero exit codes, `is_error` flags, exception keywords)
   - Repetitive tool sequences (3+ tools repeated 3+ times)
   - Automation request keywords (English and Japanese)
   - Unresolved requests (5+ minute gap after user message)

…

Evidence boundary and execution chain

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

sections -> When to Use → Prerequisites → Workflow → Quick Start → Stage 1: Session Log Mining → Stage 2: Scoring and Deduplication
terms    -> Python 3.10+ · Claude CLI · Session logs · 1. Enumerate session logs from allowlist projects in ~/.claude/projects/ 2. · 1. Load existing skills from skills//SKILL.md frontmatter 2.
files/cmd -> python3 scripts/runskillgenerationpipeline.py --mode weekly · pyyaml · claude --version · ~/.claude/projects/<project>/ · ~/.claude/projects/ · timestamp · type: "user" · userType: "external"
body sha256 -> 2f6a8e4ffe61